Skip to content
Permalink
Branch: master
Find file Copy path
Find file Copy path
4038 lines (2832 sloc) 169 KB

Flux Specification

The following document specifies the Flux language and query execution.

This document is a living document and does not represent the current implementation of Flux. Any section that is not currently implemented is commented with a IMPL#XXX where XXX is an issue number tracking discussion and progress towards implementation.

Language

The Flux language is centered on querying and manipulating time series data.

Notation

The syntax of the language is specified using Extended Backus-Naur Form (EBNF):

Production  = production_name "=" [ Expression ] "." .
Expression  = Alternative { "|" Alternative } .
Alternative = Term { Term } .
Term        = production_name | token [ "…" token ] | Group | Option | Repetition .
Group       = "(" Expression ")" .
Option      = "[" Expression "]" .
Repetition  = "{" Expression "}" .

Productions are expressions constructed from terms and the following operators, in increasing precedence:

|   alternation
()  grouping
[]  option (0 or 1 times)
{}  repetition (0 to n times)

Lower-case production names are used to identify lexical tokens. Non-terminals are in CamelCase. Lexical tokens are enclosed in double quotes "" or back quotes ``.

Representation

Source code is encoded in UTF-8. The text need not be canonicalized.

Characters

This document will use the term character to refer to a Unicode code point.

The following terms are used to denote specific Unicode character classes:

newline        = /* the Unicode code point U+000A */ .
unicode_char   = /* an arbitrary Unicode code point except newline */ .
unicode_letter = /* a Unicode code point classified as "Letter" */ .
unicode_digit  = /* a Unicode code point classified as "Number, decimal digit" */ .

In The Unicode Standard 8.0, Section 4.5 "General Category" defines a set of character categories. Flux treats all characters in any of the Letter categories Lu, Ll, Lt, Lm, or Lo as Unicode letters, and those in the Number category Nd as Unicode digits.

Letters and digits

The underscore character _ (U+005F) is considered a letter.

letter        = unicode_letter | "_" .
decimal_digit = "0" … "9" .

Lexical Elements

Comments

Comment serve as documentation. Comments begin with the character sequence // and stop at the end of the line.

Comments cannot start inside string or regexp literals. Comments act like newlines.

Tokens

Flux is built up from tokens. There are several classes of tokens: identifiers, keywords, operators, and literals. White space, formed from spaces, horizontal tabs, carriage returns, and newlines, is ignored except as it separates tokens that would otherwise combine into a single token. While breaking the input into tokens, the next token is the longest sequence of characters that form a valid token.

Identifiers

Identifiers name entities within a program. An identifier is a sequence of one or more letters and digits. An identifier must start with a letter.

identifier = letter { letter | unicode_digit } .

Examples:

a
_x
longIdentifierName
αβ

Keywords

The following keywords are reserved and may not be used as identifiers:

and    import  not  return   option   test
empty  in      or   package  builtin

IMPL#256 Add in and empty operator support

Operators

The following character sequences represent operators:

+   ==   !=   (   )   =>
-   <    !~   [   ]
*   >    =~   {   }
/   <=   =    ,   :
%   >=   <-   .   |>

Numeric literals

Numeric literals may be integers or floating point values. Literals have arbitrary precision and will be coerced to a specific type when used.

The following coercion rules apply to numeric literals:

  • an integer literal can be coerced to an "int", "uint", or "float" type,
  • an float literal can be coerced to a "float" type,
  • an error will occur if the coerced type cannot represent the literal value.

IMPL#255 Allow numeric literal coercion.

Integer literals

An integer literal is a sequence of digits representing an integer value. Only decimal integers are supported.

int_lit     = "0" | decimal_lit .
decimal_lit = ( "1" … "9" ) { decimal_digit } .

Examples:

0
42
317316873
Floating-point literals

A floating-point literal is a decimal representation of a floating-point value. It has an integer part, a decimal point, and a fractional part. The integer and fractional part comprise decimal digits. One of the integer part or the fractional part may be elided.

float_lit = decimals "." [ decimals ]
          | "." decimals .
decimals  = decimal_digit { decimal_digit } .

Examples:

0.
72.40
072.40  // == 72.40
2.71828
.26

IMPL#254 Parse float literals

Duration literals

A duration literal is a representation of a length of time. It has an integer part and a duration unit part. Multiple durations may be specified together and the resulting duration is the sum of each smaller part. When several durations are specified together, larger units must appear before smaller ones, and there can be no repeated units.

duration_lit  = { int_lit duration_unit } .
duration_unit = "y" | "mo" | "w" | "d" | "h" | "m" | "s" | "ms" | "us" | "µs" | "ns" .
Units Meaning
y year (12 months)
mo month
w week (7 days)
d day
h hour (60 minutes)
m minute (60 seconds)
s second
ms milliseconds (1 thousandth of a second)
us or µs microseconds (1 millionth of a second)
ns nanoseconds (1 billionth of a second)

Durations represent a length of time. Lengths of time are dependent on specific instants in time they occur and as such, durations do not represent a fixed amount of time. No amount of seconds is equal to a day, as days vary in their number of seconds. No amount of days is equal to a month, as months vary in their number of days. A duration consists of three basic time units: seconds, days and months.

Durations can be combined via addition and subtraction. Durations can be multiplied by an integer value. These operations are performed on each time unit independently.

Examples:

1s
10d
1h15m // 1 hour and 15 minutes
5w
1mo5d // 1 month and 5 days

Durations can be added to date times to produce a new date time.

Addition and subtraction of durations to date times do not commute and are left associative. Addition and subtraction of durations to date times applies months, days and seconds in that order. When months are added to a date times and the resulting date is past the end of the month, the day is rolled back to the last day of the month.

Examples:

2018-01-01T00:00:00Z + 1d       // 2018-01-02T00:00:00Z
2018-01-01T00:00:00Z + 1mo      // 2018-02-01T00:00:00Z
2018-01-01T00:00:00Z + 2mo      // 2018-03-01T00:00:00Z
2018-01-31T00:00:00Z + 2mo      // 2018-03-31T00:00:00Z
2018-02-28T00:00:00Z + 2mo      // 2018-04-28T00:00:00Z
2018-01-31T00:00:00Z + 1mo      // 2018-02-28T00:00:00Z, February 31th is rolled back to the last day of the month, February 28th in 2018.

// Addition and subtraction of durations to date times does not commute
2018-02-28T00:00:00Z + 1mo + 1d // 2018-03-29T00:00:00Z
2018-02-28T00:00:00Z + 1d + 1mo // 2018-04-01T00:00:00Z
2018-01-01T00:00:00Z + 2mo - 1d // 2018-02-28T00:00:00Z
2018-01-01T00:00:00Z - 1d + 3mo // 2018-03-31T00:00:00Z

// Addition and subtraction of durations to date times applies months, days and seconds in that order.
2018-01-28T00:00:00Z + 1mo + 2d // 2018-03-02T00:00:00Z
2018-01-28T00:00:00Z + 1mo2d    // 2018-03-02T00:00:00Z
2018-01-28T00:00:00Z + 2d + 1mo // 2018-02-28T00:00:00Z, explicit left associative add of 2d first changes the result
2018-02-01T00:00:00Z + 2mo2d    // 2018-04-03T00:00:00Z
2018-01-01T00:00:00Z + 1mo30d   // 2018-03-02T00:00:00Z, Months are applied first to get February 1st, then days are added resulting in March 2 in 2018.
2018-01-31T00:00:00Z + 1mo1d    // 2018-03-01T00:00:00Z, Months are applied first to get February 28th, then days are added resulting in March 1 in 2018.

IMPL#657 Implement Duration vectors

Date and time literals

A date and time literal represents a specific moment in time. It has a date part, a time part and a time offset part. The format follows the RFC 3339 specification. The time is optional, when it is omitted the time is assumed to be midnight for the default location. The time_offset is optional, when it is omitted the location option is used to determine the offset.

date_time_lit     = date [ "T" time ] .
date              = year_lit "-" month "-" day .
year              = decimal_digit decimal_digit decimal_digit decimal_digit .
month             = decimal_digit decimal_digit .
day               = decimal_digit decimal_digit .
time              = hour ":" minute ":" second [ fractional_second ] [ time_offset ] .
hour              = decimal_digit decimal_digit .
minute            = decimal_digit decimal_digit .
second            = decimal_digit decimal_digit .
fractional_second = "."  { decimal_digit } .
time_offset       = "Z" | ("+" | "-" ) hour ":" minute .

Examples:

1952-01-25T12:35:51Z
2018-08-15T13:36:23-07:00
2009-10-15T09:00:00       // October 15th 2009 at 9 AM in the default location
2018-01-01                // midnight on January 1st 2018 in the default location

IMPL#152 Implement shorthand time literals

String literals

A string literal represents a sequence of characters enclosed in double quotes. Within the quotes any character may appear except an unescaped double quote. String literals support several escape sequences.

\n   U+000A line feed or newline
\r   U+000D carriage return
\t   U+0009 horizontal tab
\"   U+0022 double quote
\\   U+005C backslash
\{   U+007B open curly bracket
\}   U+007D close curly bracket

Additionally any byte value may be specified via a hex encoding using \x as the prefix.

string_lit       = `"` { unicode_value | byte_value | StringExpression | newline } `"` .
byte_value       = `\` "x" hex_digit hex_digit .
hex_digit        = "0" … "9" | "A" … "F" | "a" … "f" .
unicode_value    = unicode_char | escaped_char .
escaped_char     = `\` ( "n" | "r" | "t" | `\` | `"` ) .
StringExpression = "{" Expression "}" .

TODO(nathanielc): With string interpolation string_lit is not longer a lexical token as part of a literal, but an entire expression in and of itself.

IMPL#252 Parse string literals

Examples:

"abc"
"string with double \" quote"
"string with backslash \\"
"日本語"
"\xe6\x97\xa5\xe6\x9c\xac\xe8\xaa\x9e" // the explicit UTF-8 encoding of the previous line

String literals are also interpolated for embedded expressions to be evaluated as strings. Embedded expressions are enclosed in curly brackets "{}". The expressions are evaluated in the scope containing the string literal. The result of an expression is formatted as a string and replaces the string content between the brackets. All types are formatted as strings according to their literal representation. A function "printf" exists to allow more precise control over formatting of various types. To include the literal curly brackets within a string they must be escaped.

IMPL#248 Add printf function

Interpolation example:

n = 42
"the answer is {n}" // the answer is 42
"the answer is not {n+1}" // the answer is not 43
"openinng curly bracket \{" // openinng curly bracket {
"closing curly bracket \}" // closing curly bracket }

IMPL#251 Add string interpolation support

Regular expression literals

A regular expression literal represents a regular expression pattern, enclosed in forward slashes. Within the forward slashes, any unicode character may appear except for an unescaped forward slash. The \x hex byte value representation from string literals may also be present.

Regular expression literals support only the following escape sequences:

\/   U+002f forward slash
\\   U+005c backslash


regexp_lit         = "/" regexp_char { regexp_char } "/" .
regexp_char        = unicode_char | byte_value | regexp_escape_char .
regexp_escape_char = `\` (`/` | `\`)

Examples:

/.*/
/http:\/\/localhost:9999/
/^\xe6\x97\xa5\xe6\x9c\xac\xe8\xaa\x9e(ZZ)?$/
/^日本語(ZZ)?$/ // the above two lines are equivalent
/\\xZZ/ // this becomes the literal pattern "\xZZ"
/a\/b\\c\d/ // escape sequences and character class shortcuts are supported
/(?:)/ // the empty regular expression

The regular expression syntax is defined by RE2.

Variables

A variable represents a storage location for a single value. Variables are immutable. Once a variable is given a value, it holds that value for the remainder of its lifetime.

Options

An option represents a storage location for any value of a specified type. Options are mutable. An option can hold different values during its lifetime.

Below is a list of some built-in options that are currently implemented in the Flux language:

  • now
  • task
  • location
now

The now option is a function that returns a time value to be used as a proxy for the current system time.

// Query should execute as if the below time is the current system time
option now = () => 2006-01-02T15:04:05-07:00
task

The task option is used by a scheduler to schedule the execution of a Flux query.

option task = {
    name: "foo",        // name is required
    every: 1h,          // task should be run at this interval
    delay: 10m,         // delay scheduling this task by this duration
    cron: "0 2 * * *",  // cron is a more sophisticated way to schedule. every and cron are mutually exclusive
    retry: 5,           // number of times to retry a failed query
}
location

The location option is used to set the default time zone of all times in the script. The location maps the UTC offset in use at that location for a given time. The default value is set using the time zone of the running process.

option location = fixedZone(offset:-5h) // set timezone to be 5 hours west of UTC
option location = loadLocation(name:"America/Denver") // set location to be America/Denver

IMPL#660 Implement Location option

Types

A type defines a set of values and operations on those values. Types are never explicitly declared as part of the syntax. Types are always inferred from the usage of the value. Type inference follows a Hindley-Milner style inference system.

Union types

A union type defines a set of types. In the rest of this section a union type will be specified as follows:

T = t1 | t2 | ... | tn

where t1, t2, ..., and tn are types. In the example above a value of type T is either of type t1, type t2, ..., or type tn.

Basic types

These are the types from which all other Flux data types are constructed.

Null type

The null type represents a missing or unknown value. The null type name is null. There is only one value that comprises the null type and that is the null value.

A type t is nullable if it can be expressed as follows:

t = {s} | null

where {s} defines a set of values.

Boolean types

A boolean type represents a truth value, corresponding to the preassigned variables true and false. The boolean type name is bool. The boolean type is nullable and can be formally specified as follows:

bool = {true, false} | null
Numeric types

A numeric type represents sets of integer or floating-point values.

The following numeric types exist:

uint    = {the set of all unsigned 64-bit integers} | null
int     = {the set of all signed 64-bit integers} | null
float   = {the set of all IEEE-754 64-bit floating-point numbers} | null

Note all numeric types are nullable.

Time types

A time type represents a single point in time with nanosecond precision. The time type name is time. The time type is nullable.

Duration types

A duration type represents a length of time with nanosecond precision. The duration type name is duration. The duration type is nullable.

Durations can be added to times to produce a new time.

Examples:

2018-07-01T00:00:00Z + 1mo // 2018-08-01T00:00:00Z
2018-07-01T00:00:00Z + 2y  // 2020-07-01T00:00:00Z
2018-07-01T00:00:00Z + 5h  // 2018-07-01T05:00:00Z
String types

A string type represents a possibly empty sequence of characters. Strings are immutable: once created they cannot be modified. The string type name is string. The string type is nullable. Note that an empty string is distinct from a null value.

The length of a string is its size in bytes, not the number of characters, since a single character may be multiple bytes.

Regular expression types

A regular expression type represents the set of all patterns for regular expressions. The regular expression type name is regexp. The regular expression type is not nullable.

Composite types

These are types constructed from basic types. Composite types are not nullable.

Array types

An array type represents a sequence of values of any other type. All values in the array must be of the same type. The length of an array is the number of elements in the array.

Object types

An object type represents a set of unordered key and value pairs. The key must always be a string. The value may be any other type, and need not be the same as other values within the object.

Function types

A function type represents a set of all functions with the same argument and result types.

IMPL#249 Specify type inference rules

Generator types

A generator type represents a value that produces an unknown number of other values. The generated values may be of any other type but must all be the same type.

IMPL#658 Implement Generators types

Polymorphism

Flux types can be polymorphic, meaning that a type may take on many different types. Flux supports let-polymorphism and structural polymorphism.

Let-polymorphism is the concept that each time an identifier is referenced is may take on a different type. For example:

add = (a,b) => a + b
add(a:1,b:2) // 3
add(a:1.5,b:2.0) // 3.5

The identifiers a and b in the body of the add function are used as both int and float types. This is let-polymorphism, each different use of an identifier may have a different type.

Structural polymorphism is the concept that structures (objects in Flux) can be used by the same function even if the structures themselves are different. For example:

john = {name:"John", lastName:"Smith"}
jane = {name:"Jane", age:44}

// John and Jane are objects with different types.
// We can still define a function that can operate on both objects safely.

// name returns the name of a person
name = (person) => person.name

name(person:john) // John
name(person:jane) // Jane

device = {id: 125325, lat: 15.6163, lon: 62.6623}

name(person:device) // Type error, "device" does not have a property name.

This is structural polymorphism, objects of differing types can be used as the same type so long as they both contain the necessary properties. The necessary properties are determined by the use of the object.

This form of polymorphism means that these checks are performed during type inference and not during runtime. Type errors are found and reported before runtime.

Blocks

A block is a possibly empty sequence of statements within matching brace brackets.

Block         = "{" StatementList "} .
StatementList = { Statement } .

In addition to explicit blocks in the source code, there are implicit blocks:

  1. The universe block encompasses all Flux source text.
  2. Each package has a package block containing all Flux source text for that package.
  3. Each file has a file block containing all Flux source text in that file.
  4. Each function literal has its own function block even if not explicitly declared.

Blocks nest and influence scoping.

Assignment and scope

An assignment binds an identifier to a variable, option, or function. Every identifier in a program must be assigned.

Flux is lexically scoped using blocks:

  1. The scope of a preassigned identifier is in the universe block.
  2. The scope of an identifier denoting a variable, option, or function at the top level (outside any function) is the package block.
  3. The scope of the name of an imported package is the file block of the file containing the import declaration.
  4. The scope of an identifier denoting a function argument is the function body.
  5. The scope of an identifier assigned inside a function is the innermost containing block.

Note that the package clause is not an assignment. The package name does not appear in any scope. Its purpose is to identify the files belonging to the same package and to specify the default package name for import declarations.

IMPL#247 Add package/namespace support

Variable assignment

VariableAssignment = identifier "=" Expression

A variable assignment creates a variable bound to an identifier and gives it a type and value. A variable keeps the same type and value for the remainder of its lifetime. An identifier assigned to a variable in a block cannot be reassigned in the same block. An identifier can be reassigned or shadowed in an inner block.

Examples:

n = 1
m = 2
x = 5.4
f = () => {
    n = "a"
    m = "b"
    return a + b
}

Option assignment

OptionAssignment = "option" [ identifier "." ] identifier "=" Expression

An option assignment creates an option bound to an identifier and gives it a type and a value. Options may only be assigned in a package block. Once declared, an option may not be redeclared in the same package block. An option declared in one package may be reassigned a new value in another. An option keeps the same type for the remainder of its lifetime.

Examples:

// alert package
option severity = ["low", "moderate", "high"]

// foo package
import "alert"

option alert.severity = ["low", "critical"]  // qualified option

option n = 2

f = (a, b) => a + b + n

x = f(a:1, b:1) // x = 4

Expressions

An expression specifies the computation of a value by applying the operators and functions to operands.

Operands and primary expressions

Operands denote the elementary values in an expression.

Primary expressions are the operands for unary and binary expressions. A primary expressions may be a literal, an identifier denoting a variable, or a parenthesized expression.

PrimaryExpression = identifier | Literal | "(" Expression ")" .

Logical Operators

Flux provides the logical operators and and or.
Flux's logical operators observe the short-circuiting behavior seen in other programming languages, meaning that the right-hand side (RHS) operand is conditionally evaluated depending on the result of evaluating the left-hand side (LHS) operand.

When the operator is and:

  • If the LHS operand evaluates to false, a value of false is produced and the RHS operand is not evaluated.

When the operator is or:

  • If the LHS operand evaluates to true, a value of true is produced and the RHS operand is not evaluated.

Literals

Literals construct a value.

Literal = int_lit
        | float_lit
        | string_lit
        | regex_lit
        | duration_lit
        | pipe_receive_lit
        | ObjectLiteral
        | ArrayLiteral
        | FunctionLiteral .
Object literals

Object literals construct a value with the object type.

ObjectLiteral  = "{" ObjectBody "}" .
ObjectBody     = WithProperties | PropertyList .
WithProperties = identifier "with"  PropertyList .
PropertyList   = [ Property { "," Property } ] .
Property       = identifier [ ":" Expression ]
               | string_lit ":" Expression .

Examples:

{a: 1, b: 2, c: 3}
{a, b, c}
{o with x: 5, y: 5}
{o with a, b}
Array literals

Array literals construct a value with the array type.

ArrayLiteral   = "[" ExpressionList "]" .
ExpressionList = [ Expression { "," Expression } ] .
Function literals

A function literal defines a new function with a body and parameters. The function body may be a block or a single expression. The function body must have a return statement if it is an explicit block, otherwise the expression is the return value.

FunctionLiteral    = FunctionParameters "=>" FunctionBody .
FunctionParameters = "(" [ ParameterList [ "," ] ] ")" .
ParameterList      = Parameter { "," Parameter } .
Parameter          = identifier [ "=" Expression ] .
FunctionBody       = Expression | Block .

Examples:

() => 1 // function returns the value 1
(a, b) => a + b // function returns the sum of a and b
(x=1, y=1) => x * y // function with default values
(a, b, c) => { // function with a block body
    d = a + b
    return d / c
}

All function literals are anonymous. A function may be given a name using a variable assignment.

add = (a,b) => a + b
mul = (a,b) => a * b

Function literals are closures: they may refer to variables defined is a surrounding block. Those variables are shared between the function literal and the surrounding block.

Call expressions

A call expressions invokes a function with the provided arguments. Arguments must be specified using the argument name, positional arguments not supported. Argument order does not matter. When an argument has a default value, it is not required to be specified.

CallExpression = "(" PropertyList ")" .

Examples:

f(a:1, b:9.6)
float(v:1)

Pipe expressions

A pipe expression is a call expression with an implicit piped argument. Pipe expressions simplify creating long nested call chains.

Pipe expressions pass the result of the left hand expression as the pipe argument to the right hand call expression. Function literals specify which if any argument is the pipe argument using the pipe literal as the argument's default value. It is an error to use a pipe expression if the function does not declare a pipe argument.

pipe_receive_lit = "<-" .

Examples:

foo = () => // function body elided
bar = (x=<-) => // function body elided
baz = (y=<-) => // function body elided
foo() |> bar() |> baz() // equivalent to baz(x:bar(y:foo()))

Index expressions

Index expressions access a value from an array based on a numeric index.

IndexExpression = "[" Expression "]" .

Member expressions

Member expressions access a property of an object. They are specified via an expression of the form obj.k or obj["k"]. The property being accessed must be either an identifer or a string literal. In either case the literal value is the name of the property being accessed, the identifer is not evaluated. It is not possible to access an object's property using an arbitrary expression.

If obj contains an entry with property k, both obj.k and obj["k"] return the value associated with k. If obj does not contain an entry with property k, both obj.k and obj["k"] return null.

MemberExpression        = DotExpression  | MemberBracketExpression
DotExpression           = "." identifer
MemberBracketExpression = "[" string_lit "]" .

Conditional Expressions

Conditional expressions evaluate a boolean-valued condition and if the result is true, the expression following the then keyword is evaluated and returned. Otherwise the expression following the else keyword is evaluated and returned. In either case, the branch not taken is not evaluated; only side effects associated with the branch that is taken will occur.

ConditionalExpression   = "if" Expression "then" Expression "else" Expression .

Example:

color = if code == 0 then "green" else if code == 1 then "yellow" else "red"

Note according to the above definition, if a condition evaluates to a null or unknown value, the else branch is evaluated.

Operators

Operators combine operands into expressions. The precedence of the operators is given in the table below. Operators with a lower number have higher precedence.

Precedence Operator Description
1 a() Function call
a[] Member or index access
. Member access
2 * / Multiplication and division
3 + - Addition and subtraction
4 == != Comparison operators
< <=
> >=
=~ !~
5 not Unary logical operator
exists Null check operator
6 and Logical AND
7 or Logical OR
8 if/then/else Conditional

The operator precedence is encoded directly into the grammar as the following.

Expression               = ConditionalExpression .
ConditionalExpression    = LogicalExpression
                         | "if" Expression "then" Expression "else" Expression .
LogicalExpression        = UnaryLogicalExpression
                         | LogicalExpression LogicalOperator UnaryLogicalExpression .
LogicalOperator          = "and" | "or" .
UnaryLogicalExpression   = ComparisonExpression
                         | UnaryLogicalOperator UnaryLogicalExpression .
UnaryLogicalOperator     = "not" | "exists" .
ComparisonExpression     = MultiplicativeExpression
                         | ComparisonExpression ComparisonOperator MultiplicativeExpression .
ComparisonOperator       = "==" | "!=" | "<" | "<=" | ">" | ">=" | "=~" | "!~" .
AdditiveExpression       = MultiplicativeExpression
                         | AdditiveExpression AdditiveOperator MultiplicativeExpression .
AdditiveOperator         = "+" | "-" .
MultiplicativeExpression = PipeExpression
                         | MultiplicativeExpression MultiplicativeOperator PipeExpression .
MultiplicativeOperator   = "*" | "/" .
PipeExpression           = PostfixExpression
                         | PipeExpression PipeOperator UnaryExpression .
PipeOperator             = "|>" .
UnaryExpression          = PostfixExpression
                         | PrefixOperator UnaryExpression .
PrefixOperator           = "+" | "-" .
PostfixExpression        = PrimaryExpression
                         | PostfixExpression PostfixOperator .
PostfixOperator          = MemberExpression
                         | CallExpression
                         | IndexExpression .

Packages

Flux source is organized into packages. A package consists of one or more source files. Each source file is parsed individually and composed into a single package.

File = [ PackageClause ] [ ImportList ] StatementList .
ImportList = { ImportDeclaration } .

Package clause

PackageClause = "package" identifier .

A package clause defines the name for the current package. Package names must be valid Flux identifiers. The package clause must be at the begining of any Flux source file. All files in the same package must declare the same package name. When a file does not declare a package clause, all identifiers in that file will belong to the special main package.

IMPL#247 Add package/namespace support

package main

The main package is special for a few reasons:

  1. It defines the entrypoint of a Flux program
  2. It cannot be imported
  3. All statements are marked as producing side effects

Statements

A statement controls execution.

Statement = OptionAssignment
          | BuiltinStatement
          | VariableAssignment
          | ReturnStatement
          | ExpressionStatement .

Import declaration

ImportDeclaration = "import" [identifier] string_lit

Associated with every package is a package name and an import path. The import statement takes a package's import path and brings all of the identifiers defined in that package into the current scope under a namespace. The import statment defines the namespace through which to access the imported identifiers. By default the identifier of this namespace is the package name unless otherwise specified. For example, given a variable x declared in package foo, importing foo and referencing x would look like this:

import "import/path/to/package/foo"

foo.x

Or this:

import bar "import/path/to/package/foo"

bar.x

A package's import path is always absolute. A package may reassign a new value to an option identifier declared in one of its imported packages. A package cannot access nor modify the identifiers belonging to the imported packages of its imported packages. Every statement contained in an imported package is evaluated.

Return statements

A terminating statement prevents execution of all statements that appear after it in the same block. A return statement is a terminating statement.

ReturnStatement = "return" Expression .

Expression statements

An expression statement is an expression where the computed value is discarded.

ExpressionStatement = Expression .

Examples:

1 + 1
f()
a

Named types

A named type can be created using a type assignment statement. A named type is equivalent to the type it describes and may be used interchangeably.

TypeAssignement   = "type" identifier "=" TypeExpression
TypeExpression    = identifier
                  | TypeParameter
                  | ObjectType
                  | ArrayType
                  | GeneratorType
                  | FunctionType .
TypeParameter     = "'" identifier .
ObjectType        = "{" PropertyTypeList [";" ObjectUpperBound ] "}" .
ObjectUpperBound  = "any" | PropertyTypeList .
PropertyTypeList  = PropertyType [ "," PropertyType ] .
PropertyType      = identifier ":" TypeExpression
                  | string_lit ":" TypeExpression .
ArrayType         = "[]" TypeExpression .
GeneratorType     = "[...]" TypeExpression .
FunctionType      = ParameterTypeList "->" TypeExpression
ParameterTypeList = "(" [ ParameterType { "," ParameterType } ] ")" .
ParameterType     = identifier ":" [ pipe_receive_lit ] TypeExpression .

Named types are a separate namespace from values. It is possible for a value and a type to have the same identifier. The following named types are built-in.

bool     // boolean
int      // integer
uint     // unsigned integer
float    // floating point number
duration // duration of time
time     // time
string   // utf-8 encoded string
regexp   // regular expression
type     // a type that itself describes a type

When an object's upper bound is not specified, it is assumed to be equal to its lower bound.

Parameters to function types define whether the parameter is a pipe forward parameter and whether the parameter has a default value. The <- indicates the parameter is the pipe forward parameter.

Examples:

// alias the bool type
type boolean = bool

// define a person as an object type
type person = {
    name: string,
    age: int,
}

// Define addition on ints
type intAdd = (a: int, b: int) -> int

// Define polymorphic addition
type add = (a: 'a, b: 'a) -> 'a

// Define funcion with pipe parameter
type bar = (foo: <-string) -> string

// Define object type with an empty lower bound and an explicit upper bound
type address = {
    ;
    street: string,
    city: string,
    state: string,
    country: string,
    province: string,
    zip: int,
}

Side Effects

Side effects can occur in two ways.

  1. By reassigning builtin options
  2. By calling a function that produces side effects

A function produces side effects when it is explicitly declared to have side effects or when it calls a function that itself produces side effects.

Package initialization

Packages are initialized in the following order:

  1. All imported packages are initialized and assigned to their package identifier.
  2. All option declarations are evaluated and assigned regardless of order. An option cannot have a dependency on another option assigned in the same package block.
  3. All variable declarations are evaluated and assigned regardless of order. A variable cannot have a direct or indirect dependency on itself.
  4. Any package side effects are evaluated.

A package will only be initialized once across all file blocks and across all packages blocks regardless of how many times it is imported.

Initializing imported packages must be deterministic. Specifically after all imported packages are initialized, each option must be assigned the same value. Packages imported in the same file block are initialized in declaration order. Packages imported across different file blocks have no known order. When a set of imports modify the same option, they must be ordered by placing them in the same file block.

Built-ins

Flux contains many preassigned values. These preassigned values are defined in the source files for the various built-in packages.

System built-ins

When a built-in value is not expressible in Flux its value may be defined by the hosting environment. All such values must have a corresponding builtin statement to declare the existence and type of the built-in value.

BuiltinStatement = "builtin" identifer ":" TypeExpression

Example

builtin from : (bucket: string, bucketID: string) -> stream

Time constants

Days of the week

Days of the week are represented as integers in the range [0-6]. The following builtin values are defined:

Sunday    = 0
Monday    = 1
Tuesday   = 2
Wednesday = 3
Thursday  = 4
Friday    = 5
Saturday  = 6

IMPL#153 Add Days of the Week constants

Months of the year

Months are represented as integers in the range [1-12]. The following builtin values are defined:

January   = 1
February  = 2
March     = 3
April     = 4
May       = 5
June      = 6
July      = 7
August    = 8
September = 9
October   = 10
November  = 11
December  = 12

IMPL#154 Add Months of the Year constants

Time and date functions

These are builtin functions that all take a single time argument and return an integer.

  • second int Second returns the second of the minute for the provided time in the range [0-59].
  • minute int Minute returns the minute of the hour for the provided time in the range [0-59].
  • hour int Hour returns the hour of the day for the provided time in the range [0-59].
  • weekDay int WeekDay returns the day of the week for the provided time in the range [0-6].
  • monthDay int MonthDay returns the day of the month for the provided time in the range [1-31].
  • yearDay int YearDay returns the day of the year for the provided time in the range [1-366].
  • month int Month returns the month of the year for the provided time in the range [1-12].

IMPL#155 Implement Time and date functions

System Time

The builtin function systemTime returns the current system time. All calls to systemTime within a single evaluation of a Flux script return the same time.

Intervals

Intervals is a function that produces a set of time intervals over a range of time. An interval is an object with start and stop properties that correspond to the inclusive start and exclusive stop times of the time interval. The return value of intervals is another function that accepts start and stop time parameters and returns an interval generator. The generator is then used to produce the set of intervals. The set of intervals will include all intervals that intersect with the initial range of time. The intervals function is designed to be used with the intervals parameter of the window function.

By default the end boundary of an interval will align with the Unix epoch (zero time) modified by the offset of the location option.

An interval is a built-in named type:

type interval = {
    start: time,
    stop: time,
}

Intervals has the following parameters:

Name Type Description
every duration Every is the duration between starts of each of the intervals. Defaults to the value of the period duration.
period duration Period is the length of each interval. It can be negative, indicating the start and stop boundaries are reversed. Defaults to the value of the every duration.
offset duration Offset is the duration by which to shift the window boundaries. It can be negative, indicating that the offset goes backwards in time. Defaults to 0, which will align window end boundaries with the every duration.
filter (interval: interval) -> bool Filter accepts an interval object and returns a boolean value. Defaults to include all intervals.

The Nth interval start date is the initial start date plus the offset plus an Nth multiple of the every parameter. Each interval stop date is equal to the interval start date plus the period duration. When filtering intervals each potential interval is passed to the filter function, when the function returns false, that interval is excluded from the set of intervals.

The intervals function has the following signature:

(start: time, stop: time) -> (start: time, stop: time) -> [...]interval

Examples:

intervals(every:1h)                        // 1 hour intervals
intervals(every:1h, period:2h)             // 2 hour long intervals every 1 hour
intervals(every:1h, period:2h, offset:30m) // 2 hour long intervals every 1 hour starting at 30m past the hour
intervals(every:1w, offset:1d)             // 1 week intervals starting on Monday (by default weeks start on Sunday)
intervals(every:1d, period:-1h)            // the hour from 11PM - 12AM every night
intervals(every:1mo, period:-1d)           // the last day of each month

Examples using a predicate:

// 1 day intervals excluding weekends
intervals(
    every:1d,
    filter: (interval) => !(weekday(time: interval.start) in [Sunday, Saturday]),
)
// Work hours from 9AM - 5PM on work days.
intervals(
    every:1d,
    period:8h,
    offset:9h,
    filter:(interval) => !(weekday(time: interval.start) in [Sunday, Saturday]),
)

Examples using known start and stop dates:

// Every hour for six hours on Sep 5th.
intervals(every:1h)(start:2018-09-05T00:00:00-07:00, stop: 2018-09-05T06:00:00-07:00)
// [2018-09-05T00:00:00-07:00, 2018-09-05T01:00:00-07:00)
// [2018-09-05T01:00:00-07:00, 2018-09-05T02:00:00-07:00)
// [2018-09-05T02:00:00-07:00, 2018-09-05T03:00:00-07:00)
// [2018-09-05T03:00:00-07:00, 2018-09-05T04:00:00-07:00)
// [2018-09-05T04:00:00-07:00, 2018-09-05T05:00:00-07:00)
// [2018-09-05T05:00:00-07:00, 2018-09-05T06:00:00-07:00)

// Every hour for six hours with 1h30m periods on Sep 5th
intervals(every:1h, period:1h30m)(start:2018-09-05T00:00:00-07:00, stop: 2018-09-05T06:00:00-07:00)
// [2018-09-05T00:00:00-07:00, 2018-09-05T01:30:00-07:00)
// [2018-09-05T01:00:00-07:00, 2018-09-05T02:30:00-07:00)
// [2018-09-05T02:00:00-07:00, 2018-09-05T03:30:00-07:00)
// [2018-09-05T03:00:00-07:00, 2018-09-05T04:30:00-07:00)
// [2018-09-05T04:00:00-07:00, 2018-09-05T05:30:00-07:00)
// [2018-09-05T05:00:00-07:00, 2018-09-05T06:30:00-07:00)

// Every hour for six hours using the previous hour on Sep 5th
intervals(every:1h, period:-1h)(start:2018-09-05T12:00:00-07:00, stop: 2018-09-05T18:00:00-07:00)
// [2018-09-05T11:00:00-07:00, 2018-09-05T12:00:00-07:00)
// [2018-09-05T12:00:00-07:00, 2018-09-05T13:00:00-07:00)
// [2018-09-05T13:00:00-07:00, 2018-09-05T14:00:00-07:00)
// [2018-09-05T14:00:00-07:00, 2018-09-05T15:00:00-07:00)
// [2018-09-05T15:00:00-07:00, 2018-09-05T16:00:00-07:00)
// [2018-09-05T16:00:00-07:00, 2018-09-05T17:00:00-07:00)
// [2018-09-05T17:00:00-07:00, 2018-09-05T18:00:00-07:00)

// Every month for 4 months starting on Jan 1st
intervals(every:1mo)(start:2018-01-01, stop: 2018-05-01)
// [2018-01-01, 2018-02-01)
// [2018-02-01, 2018-03-01)
// [2018-03-01, 2018-04-01)
// [2018-04-01, 2018-05-01)

// Every month for 4 months starting on Jan 15th
intervals(every:1mo)(start:2018-01-15, stop: 2018-05-15)
// [2018-01-15, 2018-02-15)
// [2018-02-15, 2018-03-15)
// [2018-03-15, 2018-04-15)
// [2018-04-15, 2018-05-15)

IMPL#659 Implement intervals function

Builtin Intervals

The following builtin intervals exist:

// 1 second intervals
seconds = intervals(every:1s)
// 1 minute intervals
minutes = intervals(every:1m)
// 1 hour intervals
hours = intervals(every:1h)
// 1 day intervals
days = intervals(every:1d)
// 1 day intervals excluding Sundays and Saturdays
weekdays = intervals(every:1d, filter: (interval) => weekday(time:interval.start) not in [Sunday, Saturday])
// 1 day intervals including only Sundays and Saturdays
weekdends = intervals(every:1d, filter: (interval) => weekday(time:interval.start) in [Sunday, Saturday])
// 1 week intervals
weeks = intervals(every:1w)
// 1 month interval
months = intervals(every:1mo)
// 3 month intervals
quarters = intervals(every:3mo)
// 1 year intervals
years = intervals(every:1y)

FixedZone

FixedZone creates a location based on a fixed time offset from UTC.

FixedZone has the following parameters:

Name Type Description
offset duration Offset is the offset from UTC for the time zone. Offset must be less than 24h. Defaults to 0, which produces the UTC location.

Examples:

fixedZone(offset:-5h) // time zone 5 hours west of UTC
fixedZone(offset:4h30m) // time zone 4 and a half hours east of UTC

IMPL#156 Implement FixedZone function

LoadLocation

LoadLoacation loads a locations from a time zone database.

LoadLocation has the following parameters:

Name Type Description
name string Name is the name of the location to load. The names correspond to names in the IANA tzdb.

Examples:

loadLocation(name:"America/Denver")
loadLocation(name:"America/Chicago")
loadLocation(name:"Africa/Tunis")

IMPL#157 Implement LoadLoacation function

Data model

Flux employs a basic data model built from basic data types. The data model consists of tables, records, columns and streams.

Record

A record is a tuple of named values and is represented using an object type.

Column

A column has a label and a data type.

The available data types for a column are:

bool     a boolean value, true or false.
uint     an unsigned 64-bit integer
int      a signed 64-bit integer
float    an IEEE-754 64-bit floating-point number
string   a sequence of unicode characters
bytes    a sequence of byte values
time     a nanosecond precision instant in time
duration a nanosecond precision duration of time

Table

A table is set of records, with a common set of columns and a group key.

The group key is a list of columns. A table's group key denotes which subset of the entire dataset is assigned to the table. As such, all records within a table will have the same values for each column that is part of the group key. These common values are referred to as the group key value, and can be represented as a set of key value pairs.

A tables schema consists of its group key, and its column's labels and types.

IMPL#463 Specify the primitive types that make up stream and table types

Stream of tables

A stream represents a potentially unbounded set of tables. A stream is grouped into individual tables using the group key. Within a stream each table's group key value is unique.

IMPL#463 Specify the primitive types that make up stream and table types

Missing values (null)

null is a predeclared identifier representing a missing or unknown value. null is the only value comprising the null type.

Any non-boolean operator that operates on basic types, returns null when at least one of its operands is null. This can be explained intuitively with the following table and by thinking of a null value as an unknown value.

Expression Evaluates To Because
null + 5 null Adding 5 to an unknown value is still unknown
null * 5 null Multiplying an unknown value by 5 is still unknown
null == 5 null We don't know if an unknown value is equal to 5
null < 5 null We don't know if an unknown value is less than 5
null == null null We don't know if something unknown is equal to something else that is also unknown

In other words, operating on something unknown produces something that is still unknown. The only place where this is not the case is in boolean logic.

Because boolean types are nullable, Flux implements ternary logic as a way of handling boolean operators with null operands. Again, by interpreting a null operand as an unknown value, we have the following definitions:

  • not null = null
  • null or false = null
  • null or true = true
  • null or null = null
  • null and false = false
  • null and true = null
  • null and null = null

And finally, because records are represented using object types, attempting to access a column whose value is unknown or missing from a record will also return null.

Note according to the definitions above, it is not possible to check whether or not an expression is null using the == and != operators as these operators will return null if any of their operands are null. In order to perform such a check, Flux provides a built-in exists operator defined as follows:

  • exists x returns false if x is null
  • exists x returns true if x is not null

Transformations

Transformations define a change to a stream. Transformations may consume an input stream and always produce a new output stream. The order of group keys for the output stream will have a stable output order based on the input stream. The specific ordering may change between releases and it would not be considered a breaking change.

Most transformations output one table for every table they receive from the input stream. Transformations that modify the group keys or values will need to regroup the tables in the output stream. A transformation produces side effects when it is constructed from a function that produces side effects.

Transformations are represented using function types.

Built-in transformations

The following functions are preassigned in the universe block. These functions each define a transformation.

From

From produces a stream of tables from the specified bucket. Each unique series is contained within its own table. Each record in the table represents a single point in the series.

The tables schema will include the following columns:

  • _time the time of the record
  • _value the value of the record
  • _start the inclusive lower time bound of all records
  • _stop the exclusive upper time bound of all records

Additionally any tags on the series will be added as columns. The default group key for the tables is every column except _time and _value.

From has the following properties:

Name Type Description
bucket string Bucket is the name of the bucket to query.
bucketID string BucketID is the string encoding of the ID of the bucket to query.

Example:

from(bucket:"telegraf/autogen")
from(bucketID:"0261d8287f4d6000")

Buckets

Buckets is a type of data source that retrieves a list of buckets that the caller is authorized to access.
It takes no input parameters and produces an output table with the following columns:

Name Type Description
name string The name of the bucket.
id string The internal ID of the bucket.
organization string The organization this bucket belongs to.
organizationID string The internal ID of the organization.
retentionPolicy string The name of the retention policy, if present.
retentionPeriod duration The duration of time for which data is held in this bucket.

Example:

buckets() |> filter(fn: (r) => r.organization == "my-org")

Yield

Yield indicates that the stream received by the yield operation should be delivered as a result of the query. A query may have multiple results, each identified by the name provided to yield.

Yield outputs the input stream unmodified.

Yield has the following properties:

Name Type Description
name string Unique name to give to yielded results.

Example: from(bucket: "telegraf/autogen") |> range(start: -5m) |> yield(name:"1")

Note: The yield function produces side effects.

Fill

Fill will scan a stream for null values and replace them with a non-null value.

The output stream will be the same as the input stream, with all null values in the column replaced.

Fill has the following properties:

Name Type Description
column string The column to fill. Defaults to "_value"
value bool, int, uint, float, string, time The constant value to use in place of nulls. The type must match the type of the valueColumn.
usePrevious bool If set, then assign the value set in the previous non-null row. Cannot be used with value.

AssertEquals

AssertEquals is a function that will test whether two streams have identical data. It also outputs the data from the tested stream unchanged, so that this function can be used to perform in-line tests in a query.

AssertEquals has the following properties:

Name Type Description
name string Unique name given to this assertion.
got stream The stream you are testing. May be piped-forward from another function.
want stream A copy of the expected stream.

Example:

option now = () => 2015-01-01T00:00:00Z
want = from(bucket: "backup-telegraf/autogen") |> range(start: -5m)
// in-line assertion
from(bucket: "telegraf/autogen") |> range(start: -5m) |> assertEquals(want: want)

// equivalent:
got = from(bucket: "telegraf/autogen") |> range(start: -5m)
assertEquals(got: got, want: want)

Diff

Diff is a function that will produce a diff between two table streams.

It will match tables from each stream that have the same group key. For each matched table, it will produce a diff. Any rows that were added or removed will be added to the table as a row. An additional string column with the name _diff will be created which will contain a "-" if the row was present in the got table and not in the want table or "+" if the opposite is true.

The diff function is guaranteed to emit at least one row if the tables are different and no rows if the tables are the same. The exact diff that is produced may change.

Diff has the following properties:

Name Type Description
got stream The stream you are testing. May be piped-forward from another function.
want stream A copy of the expected stream.
option now = () => 2015-01-01T00:00:00Z
want = from(bucket: "backup-telegraf/autogen") |> range(start: -5m)
// in-line assertion
from(bucket: "telegraf/autogen") |> range(start: -5m) |> assertEquals(want: want)

// equivalent:
got = from(bucket: "telegraf/autogen") |> range(start: -5m)
diff(got: got, want: want)

Aggregate operations

Aggregate operations output a table for every input table they receive. A list of columns to aggregate must be provided to the operation. The aggregate function is applied to each column in isolation.

Any output table will have the following properties:

  • It always contains a single record.
  • It will have the same group key as the input table.
  • It will contain a column for each provided aggregate column. The column label will be the same as the input table. The type of the column depends on the specific aggregate operation. The value of the column will be null if the input table is empty or the input column has only null values.
  • It will not have a _time column

All aggregate operations have the following properties:

Name Type Description
column string Column specifies a column to aggregate.

The default behavior of aggregates is to skip over null values. An arbitrary aggregate function fn is expressed logically using the reduce function:

fn = (column, tables=<-) => reduce(fn: (r, accumulator) => {
    return if exists(r.column) then ... else ...
}, identity: ...)

For example, the sum transformation is logically equivalent to:

sum = (column, tables=<-) => reduce(fn: (r, accumulator) => {
    return if exists(r.column) then accumulator + r.column else accumulator
}, identity: 0)
AggregateWindow

AggregateWindow is a function that simplifies aggregating data over fixed windows of time. AggregateWindow windows the data, performs an aggregate operation, and then undoes the windowing to produce an output table for every input table.

AggregateWindow has the following properties:

Name Type Description
every duration Every specifies the window size to aggregate.
fn (tables: <-stream, columns: []string) -> stream Fn specifies the aggregate operation to perform. Any of the functions in this Aggregate section that accept a singular column parameter can be provided.
column string Columns specifies column to aggregate. Defaults to "_value".
timeSrc string TimeSrc is the name of a column from the group key to use as the source for the aggregated time. Defaults to "_stop".
timeDst string TimeDst is the name of a new column in which the aggregated time is placed. Defaults to "_time".
createEmpty bool CreateEmpty, if true, will create empty windows and fill them with a null aggregate value. Defaults to true.
Example:
// Compute the mean over 1m intervals for the last 1h.
from(bucket: "telegraf/autogen")
  |> range(start: -1h)
  |> filter(fn: (r) => r._measurement == "cpu" and r._field == "usage_system")
  |> aggregateWindow(every: 1m, fn:mean)
Covariance

Covariance is an aggregate operation. Covariance computes the covariance between two columns.

Covariance has the following properties:

Name Type Description
columns []string Columns specifies a list of the two columns to aggregate. This property is required and has no default.
pearsonr bool Pearsonr indicates whether the result should be normalized to be the Pearson R coefficient.
valueDst string ValueDst is the column into which the result will be placed. Defaults to _value.

Additionally exactly two columns must be provided to the columns property.

Example: from(bucket: "telegraf/autogen") |> range(start:-5m) |> covariance(columns: ["x", "y"])

Cov

Cov computes the covariance between two streams by first joining the streams and then performing the covariance operation per joined table.

Cov has the following properties:

Name Type Description
x stream X is one of the input streams.
y stream Y is one of the input streams.
on []string On is the list of columns on which to join.
pearsonr bool Pearsonr indicates whether the result should be normalized to be the Pearson R coefficient.
valueDst string ValueDst is the column into which the result will be placed. Defaults to _value.

Example:

cpu = from(bucket: "telegraf/autogen") |> range(start:-5m) |> filter(fn:(r) => r._measurement == "cpu")
mem = from(bucket: "telegraf/autogen") |> range(start:-5m) |> filter(fn:(r) => r._measurement == "mem")
cov(x: cpu, y: mem)

Pearsonr

Pearsonr computes the Pearson R correlation coefficient bewteen two streams. It is defined in terms of the cov function:

pearsonr = (x,y,on) => cov(x:x, y:y, on:on, pearsonr:true)
Count

Count is an aggregate operation. For each aggregated column, it outputs the number of records as an integer. It will count both null and non-null records.

Count has the following property:

Name Type Description
column string Columns specifies a column to aggregate. Defaults to "_value".

Example:

from(bucket: "telegraf/autogen") |> range(start: -5m) |> count()
Integral

Integral is an aggregate operation. For each aggregate column, it outputs the area under the curve of records. The curve is defined as a function where the domain is the record times and the range is the record values. This function will return an error if values in the time column are null or not sorted in ascending order. Null values in aggregate columns are skipped.

Integral has the following properties:

Name Type Description
column string Columns specifies a column to aggregate. Defaults to "_value".
unit duration Unit is the time duration to use when computing the integral. Defaults to 1s
timeColumn string TimeColumn is the name of the column containing the time value. Defaults to _time.

Example:

from(bucket: "telegraf/autogen")
    |> range(start: -5m)
    |> filter(fn: (r) => r._measurement == "cpu" and r._field == "usage_system")
    |> integral(unit:10s)
Mean

Mean is an aggregate operation. For each aggregated column, it outputs the mean of the non null records as a float.

Mean has the following property:

Name Type Description
column string Columns specifies a column to aggregate. Defaults to "_value".

Example:

from(bucket:"telegraf/autogen")
    |> filter(fn: (r) => r._measurement == "mem" AND
            r._field == "used_percent")
    |> range(start:-12h)
    |> window(every:10m)
    |> mean()
Median (aggregate)

Median is defined as:

median = (method="estimate_tdigest", compression=0.0, column="_value", tables=<-) =>
    tables
        |> quantile(q:0.5, method: method, compression: compression, column: column)

Is it simply a quantile with the q paramter always set to 0.5. It therefore shares all the same properties as the quantile function.

Example:

// Determine median cpu system usage:
from(bucket: "telegraf/autogen")
	|> range(start: -5m)
	|> filter(fn: (r) => r._measurement == "cpu" and r._field == "usage_system")
	|> median()
Quantile (aggregate)

Quantile is both an aggregate operation and a selector operation depending on selected options. In the aggregate methods, it outputs the value that represents the specified quantile of the non null record as a float.

Quantile has the following properties:

Name Type Description
column string Column specifies a column to aggregate. Defaults to "_value"
q float q is a value between 0 and 1 indicating the desired quantile.
method string Method must be one of: estimate_tdigest, exact_mean, or exact_selector.
compression float Compression indicates how many centroids to use when compressing the dataset. A larger number produces a more accurate result at the cost of increased memory requirements. Defaults to 1000.

The method parameter must be one of:

  • estimate_tdigest: an aggregate result that uses a tdigest data structure to compute an accurate quantile estimate on large data sources.
  • exact_mean: an aggregate result that takes the average of the two points closest to the quantile value.
  • exact_selector: see Quantile (selector)

Example:

// Determine the 0.99 quantile for cpu system usage:
from(bucket: "telegraf/autogen")
	|> range(start: -5m)
	|> filter(fn: (r) => r._measurement == "cpu" and r._field == "usage_system")
	|> quantile(q: 0.99, method: "estimate_tdigest", compression: 1000.0)
Skew

Skew is an aggregate operation. For each aggregated column, it outputs the skew of the non null record as a float.

Skew has the following parameter:

Name Type Description
column string Column specifies a columns to aggregate. Defaults to "_value".

Example:

from(bucket: "telegraf/autogen")
    |> range(start: -5m)
    |> filter(fn: (r) => r._measurement == "cpu" and r._field == "usage_system")
    |> skew()
Spread

Spread is an aggregate operation. For each aggregated column, it outputs the difference between the min and max values. The type of the output column depends on the type of input column: for input columns with type uint or int, the output is an int; for float input columns the output is a float. All other input types are invalid.

Spread has the following parameter:

Name Type Description
column string Column specifies a column to aggregate. Defaults to "_value".

Example:

from(bucket: "telegraf/autogen")
    |> range(start: -5m)
    |> filter(fn: (r) => r._measurement == "cpu" and r._field == "usage_system")
    |> spread()
Stddev

Stddev is an aggregate operation. For each aggregated column, it outputs the standard deviation of the non null record as a float.

Stddev has the following parameters:

Name Type Description
column string Columns specifies a column to aggregate. Defaults to "_value".
mode string The standard deviation mode. Must be one of "sample" or "population". Defaults to "sample".

The mode parameter specifies the type of standard deviation to calculate:

  • "sample": Calculates the sample standard deviation, where the data is considered to be part of a larger population.
  • "population": Calculates the population standard deviation, where the data is considered a population of its own.

Example:

from(bucket: "telegraf/autogen")
    |> range(start: -5m)
    |> filter(fn: (r) => r._measurement == "cpu" and r._field == "usage_system")
    |> stddev()
Sum

Stddev is an aggregate operation. For each aggregated column, it outputs the sum of the non null record. The output column type is the same as the input column type.

Sum has the following parameter:

Name Type Description
column string Column specifies a column to aggregate. Defaults to "_value".

Example:

from(bucket: "telegraf/autogen")
    |> range(start: -5m)
    |> filter(fn: (r) => r._measurement == "cpu" and r._field == "usage_system")
    |> sum()

Multiple aggregates

Multiple aggregates can be applied to the same table using the aggregate function.

IMPL#139 Add aggregate function

Selector operations

Selector operations output a table for every input table they receive. A single column on which to operate must be provided to the operation.

Any output table will have the following properties:

  • It will have the same group key as the input table.
  • It will contain the same columns as the input table.
  • It will have a column _time which represents the time of the selected record. This can be set as the value of any time column on the input table. By default the _stop time column is used.

All selector operations have the following properties:

  • column string column specifies a which column to use when selecting.
First

First is a selector operation. First selects the first non null record from the input table.

Example: from(bucket:"telegraf/autogen") |> first()

Last

Last is a selector operation. Last selects the last non null record from the input table.

Example: from(bucket: "telegraf/autogen") |> last()

Max

Max is a selector operation. Max selects the maximum record from the input table.

Example:

from(bucket:"telegraf/autogen")
    |> range(start:-12h)
    |> filter(fn: (r) => r._measurement == "cpu" AND r._field == "usage_system")
    |> max()
Min

Min is a selector operation. Min selects the minimum record from the input table.

Example:

from(bucket:"telegraf/autogen")
    |> range(start:-12h)
    |> filter(fn: (r) => r._measurement == "cpu" AND r._field == "usage_system")
    |> min()
Quantile (selector)

Quantile is both an aggregate operation and a selector operation depending on selected options. In the aggregate methods, it outputs the value that represents the specified quantile of the non null record as a float.

Quantile has the following properties:

Name Type Description
column string Column indicates which column will be used for the quantile computation. Defaults to "_value"
q float q is a value between 0 and 1 indicating the desired quantile.
method string Method must be one of: estimate_tdigest, exact_mean, exact_selector.

The method parameter must be one of:

  • estimate_tdigest: See Quantile (Aggregate).
  • exact_mean: See Quantile (Aggregate).
  • exact_selector: a selector result that returns the data point for which at least q points are less than.

Example:

// Determine the 0.99 quantile for cpu system usage:
from(bucket: "telegraf/autogen")
	|> range(start: -5m)
	|> filter(fn: (r) => r._measurement == "cpu" and r._field == "usage_system")
	|> quantile(q: 0.99, method: "exact_selector")
Median (selector)

Median is defined as:

median = (method="estimate_tdigest", compression=0.0, column="_value", tables=<-) =>
    tables
        |> quantile(q:0.5, method: method, compression: compression, column: column)

Is it simply a quantile with the q paramter always set to 0.5. It therefore shares all the same properties as the quantile function.

Example:

// Determine median cpu system usage:
from(bucket: "telegraf/autogen")
	|> range(start: -5m)
	|> filter(fn: (r) => r._measurement == "cpu" and r._field == "usage_system")
	|> median(method: "exact_selector")
Sample

Sample is a selector operation. Sample selects a subset of the records from the input table.

The following properties define how the sample is selected.

Name Type Description
n int Sample every Nth element. Must be a positive integer.
pos int Pos is the offset from start of results to begin sampling. The pos must be less than n. If pos is less than 0, a random offset is used. Default is -1 (random offset).

Example:

from(bucket:"telegraf/autogen")
    |> filter(fn: (r) => r._measurement == "cpu" AND
               r._field == "usage_system")
    |> range(start:-1d)
    |> sample(n: 5, pos: 1)

Filter

Filter applies a predicate function to each input record. Only the records for which the predicate evaluates to true will be included in the output tables. One output table is produced for each input table. The output tables will have the same schema as their corresponding input tables.

Filter has the following properties:

Name Type Description
fn (r: record) -> bool Fn is a predicate function. Records which evaluate to true, will be included in the output tables.

Example:

from(bucket:"telegraf/autogen")
    |> range(start:-12h)
    |> filter(fn: (r) => r._measurement == "cpu" AND
                r._field == "usage_system" AND
                r.service == "app-server")

Note according to the definition, records for which the predicate evaluates to null will not be included in the output.

Highest/Lowest

There are six highest/lowest functions that compute the top or bottom N records from all tables in a stream based on a specific aggregation method.

  • highestMax - computes the top N records from all tables using the maximum of each table.
  • highestAverage - computes the top N records from all tables using the average of each table.
  • highestCurrent - computes the top N records from all tables using the last value of each table.
  • lowestMin - computes the bottom N records from all tables using the minimum of each table.
  • lowestAverage - computes the bottom N records from all tables using the average of each table.
  • lowestCurrent - computes the bottom N records from all tables using the last value of each table.

All of the highest/lowest functions take the following parameters:

Name Type Description
n int N is the number of records to select.
column string Column is the column to use when aggregating. Defaults to "_value".
groupColumns []string GroupColumns are the columns on which to group to perform the aggregation.

Histogram

Histogram approximates the cumulative distribution function of a dataset by counting data frequencies for a list of bins. A bin is defined by an upper bound where all data points that are less than or equal to the bound are counted in the bin. The bin counts are cumulative.

Each input table is converted into a single output table representing a single histogram. The output table will have a the same group key as the input table. The columns not part of the group key will be removed and an upper bound column and a count column will be added.

Histogram has the following properties:

Name Type Description
column string Column is the name of a column containing the input data values. The column type must be float. Defaults to _value.
upperBoundColumn string UpperBoundColumn is the name of the column in which to store the histogram upper bounds. Defaults to le.
countColumn string CountColumn is the name of the column in which to store the histogram counts. Defaults to _value.
bins []float Bins is a list of upper bounds to use when computing the histogram frequencies. Each element in the array should contain a float value that represents the maximum value for a bin.
normalize bool Normalize when true will convert the counts into frequencies values between 0 and 1. Normalized histograms cannot be aggregated by summing their counts. Defaults to false.

Example:

histogram(bins:linearBins(start:0.0,width:10.0,count:10))  // compute the histogram of the data using 10 bins from 0,10,20,...,100

HistogramQuantile

HistogramQuantile approximates a quantile given an histogram that approximates the cumulative distribution of the dataset. Each input table represents a single histogram. The histogram tables must have two columns, a count column and an upper bound column. The count is the number of values that are less than or equal to the upper bound value. The table can have any number of records, each representing an entry in the histogram. The counts must be monotonically increasing when sorted by upper bound. If any values in the count column or upper bound column are null, an error will be returned.

Linear interpolation between the two closest bounds is used to compute the quantile. If the either of the bounds used in interpolation are infinite, then the other finite bound is used and no interpolation is performed.

The output table will have a the same group key as the input table. The columns not part of the group key will be removed and a single value column of type float will be added. The count and upper bound columns must not be part of the group key. The value column represents the value of the desired quantile from the histogram.

HistogramQuantile has the following properties:

Name Type Description
quantile float Quantile is a value between 0 and 1 indicating the desired quantile to compute.
countColumn string CountColumn is the name of the column containing the histogram counts. The count column type must be float. Defaults to _value.
upperBoundColumn string UpperBoundColumn is the name of the column containing the histogram upper bounds. The upper bound column type must be float. Defaults to le.
valueColumn string ValueColumn is the name of the output column which will contain the computed quantile. Defaults to _value.
minValue float MinValue is the assumed minumum value of the dataset. Default to 0.

When the quantile falls below the lowest upper bound, interpolation is performed between minValue and the lowest upper bound. When minValue is equal to negative infinity, the lowest upper bound is used.

Example:

histogramQuantile(quantile:0.9)  // compute the 90th quantile using histogram data.
LinearBins

LinearBins produces a list of linearly separated floats.

LinearBins has the following properties:

Name Type Description
start float Start is the first value in the returned list.
width float Width is the distance between subsequent bin values.
count int Count is the number of bins to create.
inifinity bool Infinity when true adds an additional bin with a value of positive infinity. Defaults to true.
LogarithmicBins

LogarithmicBins produces a list of exponentially separated floats.

LogarithmicBins has the following properties:

Name Type Description
start float Start is the first value in the returned bin list.
factor float Factor is the multiplier applied to each subsequent bin.
count int Count is the number of bins to create.
inifinity bool Infinity when true adds an additional bin with a value of positive infinity. Defaults to true.

Limit

Limit caps the number of records in output tables to a fixed size n. One output table is produced for each input table. Each output table will contain the first n records after the first offset records of the input table. If the input table has less than offset + n records, all records except the first offset ones will be output.

Limit has the following properties:

Name Type Description
n int N is the maximum number of records per table to output.
offset int Offest is the number of records to skip per table before limiting to n. Defaults to 0.

Example:

from(bucket: "telegraf/autogen")
    |> range(start: -1h)
    |> limit(n: 10, offset: 1)

Map

Map applies a function to each record of the input tables. The modified records are assigned to new tables based on the group key of the input table. The output tables are the result of applying the map function to each record on the input tables.

When the output record contains a different value for the group key the record is regroup into the appropriate table. When the output record drops a column that was part of the group key that column is removed from the group key.

Map has the following properties:

Name Type Description
fn (r: record) -> record Function to apply to each record. The return value must be an object.
mergeKey bool MergeKey indicates if the record returned from fn should be merged with the group key. Defaults to true.

When merging, all columns on the group key will be added to the record giving precedence to any columns already present on the record. When not merging, only columns defined on the returned record will be present on the output records.

IMPL#816 Remove mergeKey parameter from map

Example:

from(bucket:"telegraf/autogen")
    |> filter(fn: (r) => r._measurement == "cpu" AND
                r._field == "usage_system" AND
                r.service == "app-server")
    |> range(start:-12h)
    // Square the value
    |> map(fn: (r) => r._value * r._value)

Example (creating a new table):

from(bucket:"telegraf/autogen")
    |> filter(fn: (r) => r._measurement == "cpu" AND
                r._field == "usage_system" AND
                r.service == "app-server")
    |> range(start:-12h)
    // create a new table by copying each row into a new format
    |> map(fn: (r) => ({_time: r._time, app_server: r._service}))

Reduce

Reduce aggregates records in each table according to the reducer fn. The output for each table will be the group key of the table, plus columns corresponding to each field in the reducer object.

If the reducer record contains a column with the same name as a group key column, then the group key column's value is overwritten, and the outgoing group key is changed. However, if two reduced tables write to the same destination group key, then the function will error.

Reduce has the following properties:

Name Type Description
fn (r: record, accumulator: 'a) -> 'a Function to apply to each record with a reducer object of type 'a.
identity 'a an initial value to use when creating a reducer. May be used more than once in asynchronous processing use cases.

Example (compute the sum of the value column):

from(bucket:"telegraf/autogen")
    |> filter(fn: (r) => r._measurement == "cpu" AND
                r._field == "usage_system" AND
                r.service == "app-server")
    |> range(start:-12h)
    |> reduce(fn: (r, accumulator) =>
            ({sum: r._value + accumulator.sum}), identity: {sum: 0.0}))

Example (compute the sum and count in a single reducer):

from(bucket:"telegraf/autogen")
    |> filter(fn: (r) => r._measurement == "cpu" AND
                r._field == "usage_system" AND
                r.service == "app-server")
    |> range(start:-12h)
    |> reduce(fn: (r, accumulator) =>
            ({sum: r._value + accumulator.sum, count: accumulator.count + 1.0}), identity: {sum: 0.0, count:0.0}))

Example (compute the product of all values):

from(bucket:"telegraf/autogen")
    |> filter(fn: (r) => r._measurement == "cpu" AND
                r._field == "usage_system" AND
                r.service == "app-server")
    |> range(start:-12h)
    |> reduce(fn: (r, accumulator) =>
            ({prod: r._value * accumulator.prod}), identity: {prod: 1.0}))

Range

Range filters records based on provided time bounds. Each input tables records are filtered to contain only records that exist within the time bounds. Records with a null value for their time will be filtered. Each input table's group key value is modified to fit within the time bounds. Tables where all records exists outside the time bounds are filtered entirely.

IMPL#244 Update range to default to aligned window ranges.

Range has the following properties:

Name Type Description
start time Start specifies the oldest time to be included in the results.
stop time Stop specifies the exclusive newest time to be included in the results. Defaults to the value of the now option time.
timeColumn string Name of the time column to use. Defaults to _time.
startColumn string StartColumn is the name of the column containing the start time. Defaults to _start.
stopColumn string StopColumn is the name of the column containing the stop time. Defaults to _stop.

Example:

from(bucket:"telegraf/autogen")
    |> range(start:-12h, stop: -15m)
    |> filter(fn: (r) => r._measurement == "cpu" AND
               r._field == "usage_system")

Example:

from(bucket:"telegraf/autogen")
    |> range(start:2018-05-22T23:30:00Z, stop: 2018-05-23T00:00:00Z)
    |> filter(fn: (r) => r._measurement == "cpu" AND
               r._field == "usage_system")

Rename

Rename renames specified columns in a table. There are two variants: one which takes a map of old column names to new column names, and one which takes a mapping function. If a column is renamed and is part of the group key, the column name in the group key will be updated. If a specified column is not present in a table an error will be thrown.

Rename has the following properties:

Name Type Description
columns object Columns is a map of old column names to new names. Cannot be used with fn.
fn (column: string) -> string Fn defines a function mapping between old and new column names. Cannot be used with columns.

Example usage:

Rename a single column:

from(bucket: "telegraf/autogen")
    |> range(start: -5m)
    |> rename(columns: {host: "server"})

Rename all columns using fn parameter:

from(bucket: "telegraf/autogen")
    |> range(start: -5m)
    |> rename(fn: (column) => column + "_new")

Drop

Drop excludes specified columns from a table. Columns to exclude can be specified either through a list, or a predicate function. When a dropped column is part of the group key it will also be dropped from the key. If a specified column is not present in a table an error will be thrown.

Drop has the following properties:

Name Type Description
columns []string Columns is an array of column to exclude from the resulting table. Cannot be used with fn.
fn (column: string) -> bool Fn is a predicate function, columns that evaluate to true are dropped. Cannot be used with columns.

Example Usage:

Drop a list of columns:

from(bucket: "telegraf/autogen")
	|> range(start: -5m)
	|> drop(columns: ["host", "_measurement"])

Drop columns matching a predicate:

from(bucket: "telegraf/autogen")
    |> range(start: -5m)
    |> drop(fn: (column) => column =~ /usage*/)

Keep

Keep is the inverse of drop. It returns a table containing only columns that are specified, ignoring all others. Only columns in the group key that are also specified in keep will be kept in the resulting group key. If a specified column is not present in a table an error will be thrown.

Keep has the following properties:

Name Type Description
columns []string Columns is an array of column to exclude from the resulting table. Cannot be used with fn.
fn (column: string) -> bool Fn is a predicate function, columns that evaluate to true are kept. Cannot be used with columns.

Example Usage:

Keep a list of columns:

from(bucket: "telegraf/autogen")
    |> range(start: -5m)
    |> keep(columns: ["_time", "_value"])

Keep all columns matching a predicate:

from(bucket: "telegraf/autogen")
    |> range(start: -5m)
    |> keep(fn: (column) => column =~ /inodes*/)

Duplicate

Duplicate duplicates a specified column in a table. If the specified column is not present in a table an error will be thrown. If the specified column is part of the group key, it will be duplicated, but it will not be part of the group key of the output table. If the column indicated by as does not exist, a column will be added to the table. If the column does exist, that column will be overwritten with the values specified by column.
If the as column is in the group key, there are two possible outcomes: If the column indicated by column is in the group key, then as will remain in the group key and have the same group key value as column.
If column is not part of the group key, then as is removed from the group key. Duplicate has the following properties:

Name Type Description
column string Column is the name of the column to duplicate.
as string As is the name that should be assigned to the duplicate column.

Example usage:

Duplicate column server under the name host:

from(bucket: "telegraf/autogen")
	|> range(start:-5m)
	|> filter(fn: (r) => r._measurement == "cpu")
	|> duplicate(column: "host", as: "server")

Set

Set assigns a static value to each record. The key may modify and existing column or it may add a new column to the tables. If the column that is modified is part of the group key, then the output tables will be regrouped as needed.

Set has the following properties:

Name Type Description
key string Key is the label for the column to set.
value string Value is the string value to set.

Example:

from(bucket: "telegraf/autogen") |> set(key: "mykey", value: "myvalue")

Sort

Sorts orders the records within each table. One output table is produced for each input table. The output tables will have the same schema as their corresponding input tables. When sorting, nulls will always be first. When desc: false is set, then nulls are less than every other value. When desc: true, nulls are greater than every value.

Sort has the following properties:

Name Type Description
columns []string Columns is the sort order to use; precedence from left to right. Default is ["_value"].
desc bool Desc indicates results should be sorted in descending order. Default is false.

Example:

from(bucket:"telegraf/autogen")
    |> filter(fn: (r) => r._measurement == "system" AND
               r._field == "uptime")
    |> range(start:-12h)
    |> sort(columns:["region", "host", "value"])

Group

Group groups records based on their values for specific columns. It produces tables with new group keys based on the provided properties.

Group has the following properties:

Name Type Description
columns []string Columns is a list used to calculate the new group key. Defaults to [].
mode string The grouping mode, can be one of "by" or "except". Defaults to "by".

When using "by" mode, the specified columns are the new group key. When using "except" mode, the new group key is the difference between the columns of the table under exam and columns.

Examples

By

from(bucket: "telegraf/autogen")
    |> range(start: -30m)
    |> group(columns: ["host", "_measurement"])

Or:

...
    |> group(columns: ["host", "_measurement"], mode: "by")

Records are grouped by the "host" and "_measurement" columns.
The resulting group key is ["host", "_measurement"], so a new table for every different ["host", "_measurement"] value is created.
Every table in the result contains every record for some ["host", "_measurement"] value.
Every record in some resulting table has the same value for the columns "host" and "_measurement".

Except

from(bucket: "telegraf/autogen")
    |> range(start: -30m)
    |> group(columns: ["_time"], mode: "except")

Records are grouped by the set of all columns in the table, excluding "_time".
For example, if the table has columns ["_time", "host", "_measurement", "_field", "_value"] then the group key would be ["host", "_measurement", "_field", "_value"].

Single-table grouping

from(bucket: "telegraf/autogen")
    |> range(start: -30m)
    |> group()

Records are grouped into a single table.
The group key of the resulting table is empty.

Columns

Columns lists the column labels of input tables. For each input table, it outputs a table with the same group key columns, plus a new column containing the labels of the input table's columns. Each row in an output table contains the group key value and the label of one column of the input table. So, each output table has the same number of rows as the number of columns of the input table.

Columns has the following properties:

Name Type Description
column string Column is the name of the output column to store the column labels. Defaults to _value.

Example:

from(bucket: "telegraf/autogen")
    |> range(start: -30m)
    |> columns(column: "labels")

Getting every possible column label in a single table as output:

from(bucket: "telegraf/autogen")
    |> range(start: -30m)
    |> columns()
    |> keep(columns: ["_value"])
    |> group()
    |> distinct()

Keys

Keys outputs the group key of input tables. For each input table, it outputs a table with the same group key columns, plus a _value column containing the labels of the input table's group key. Each row in an output table contains the group key value and the label of one column in the group key of the input table. So, each output table has the same number of rows as the size of the group key of the input table.

Keys has the following properties:

Name Type Description
column string Column is the name of the output column to store the group key labels. Defaults to _value.

Example:

from(bucket: "telegraf/autogen")
    |> range(start: -30m)
    |> keys(column: "keys")

Getting every possible key in a single table as output:

from(bucket: "telegraf/autogen")
    |> range(start: -30m)
    |> keys()
    |> keep(columns: ["_value"])
    |> group()
    |> distinct()

KeyValues

KeyValues outputs a table with the input table's group key, plus two columns _key and _value that correspond to unique (column, value) pairs from the input table.

KeyValues has the following properties:

Name Type Description
keyColumns []string KeyColumns is a list of columns from which values are extracted.
fn (schema: schema) -> []string Fn is a schema function that may by used instead of keyColumns to identify the set of columns.

Additional requirements:

  • Only one of keyColumns or fn may be used in a single call.
  • All columns indicated must be of the same type.
  • Each input table must have all of the columns listed by the keyColumns parameter.
from(bucket: "telegraf/autogen")
    |> range(start: -30m)
    |> filter(fn: (r) => r._measurement == "cpu")
    |> keyValues(keyColumns: ["usage_idle", "usage_user"])
from(bucket: "telegraf/autogen")
    |> range(start: -30m)
    |> filter(fn: (r) => r._measurement == "cpu")
    |> keyValues(fn: (schema) => schema.columns |> filter(fn:(v) =>  v.label =~ /usage_.*/))
filterColumns = (fn) => (schema) => schema.columns |> filter(fn:(v) => fn(column:v))

from(bucket: "telegraf/autogen")
    |> range(start: -30m)
    |> filter(fn: (r) => r._measurement == "cpu")
    |> keyValues(fn: filterColumns(fn: (column) => column.label =~ /usage_.*/))

Examples:

Given the following input table with group key ["_measurement"]:

| _time | _measurement | _value | tagA |
| ----- | ------------ | ------ | ---- |
| 00001 | "m1"         | 1      | "a"  |
| 00002 | "m1"         | 2      | "b"  |
| 00003 | "m1"         | 3      | "c"  |
| 00004 | "m1"         | 4      | "b"  |

keyValues(keyColumns: ["tagA"]) produces the following table with group key ["_measurement"]:

| _measurement | _key   | _value |
| ------------ | ------ | ------ |
| "m1"         | "tagA" | "a"    |
| "m1"         | "tagA" | "b"    |
| "m1"         | "tagA" | "c"    |

keyColumns(keyColumns: ["tagB"]) produces the following error message:

received table with columns [_time, _measurement, _value, tagA] not having key columns [tagB]

Window

Window groups records based on a time value. New columns are added to uniquely identify each window and those columns are added to the group key of the output tables.

A single input record will be placed into zero or more output tables, depending on the specific windowing function.

By default the start boundary of a window will align with the Unix epoch (zero time) modified by the offset of the location option.

Window has the following properties:

Name Type Description
every duration Every is the duration of time between windows. Defaults to period's value. One of every, period or intervals must be provided.
period duration Period is the duration of the window. Period is the length of each interval. It can be negative, indicating the start and stop boundaries are reversed. Defaults to every's value. One of every, period or intervals must be provided.
offset duration Offset is the duration by which to shift the window boundaries. It can be negative, indicating that the offset goes backwards in time. Defaults to 0, which will align window end boundaries with the every duration.
intervals (start: time, stop: time) -> [...]interval Intervals is a set of intervals to be used as the windows. One of every, period or intervals must be provided. When intervals is provided, every, period, and offset must be zero.
timeColumn string TimeColumn is the name of the time column to use. Defaults to _time.
startColumn string StartColumn is the name of the column containing the window start time. Defaults to _start.
stopColumn string StopColumn is the name of the column containing the window stop time. Defaults to _stop.
createEmpty bool CreateEmpty specifies whether empty tables should be created. Defaults to false.

Example:

from(bucket:"telegraf/autogen")
    |> range(start:-12h)
    |> window(every:10m)
    |> max()
window(every:1h) // window the data into 1 hour intervals
window(intervals: intervals(every:1d, period:8h, offset:9h)) // window the data into 8 hour intervals starting at 9AM every day.

Pivot

Pivot collects values stored vertically (column-wise) in a table and aligns them horizontally (row-wise) into logical sets.

Pivot has the following properties:

Name Type Description
rowKey []string RowKey is the list of columns used to uniquely identify a row for the output.
columnKey []string ColumnKey is the list of columns used to pivot values onto each row identified by the rowKey.
valueColumn string ValueColumn identifies the single column that contains the value to be moved around the pivot.

The group key of the resulting table will be the same as the input tables, excluding the columns found in the columnKey and valueColumn. This is because these columns are not part of the resulting output table.

Any columns in the original table that are not referenced in the rowKey or the original table's group key will be dropped.

Every input row should have a 1:1 mapping to a particular row, column pair in the output table, determined by its values for the rowKey and columnKey. In the case where more than one value is identified for the same row, column pair in the output, the last value encountered in the set of table rows is taken as the result.

The output is constructed as follows:

  • The set of columns for the new table is the rowKey unioned with the group key, but excluding the columns indicated by the columnKey and the valueColumn. A new column is added to the set of columns for each unique value identified in the input by the columnKey parameter. The label of a new column is the concatenation of the values at columnKey (if the value is null, "null" is used) using _ as a separator.
  • A new row is created for each unique value identified in the input by the rowKey parameter.
  • For each new row, values for group key columns stay the same, while values for new columns are determined from the input tables by the value in valueColumn at the row identified by the rowKey values and the new column's label. If no value is found, the value is set to null.

Example 1, align fields within each measurement that have the same timestamp:

 from(bucket:"test")
     |> range(start: 1970-01-01T00:00:00.000000000Z)
     |> pivot(rowKey:["_time"], columnKey: ["_field"], valueColumn: "_value")

Input:

_time _value _measurement _field
1970-01-01T00:00:00.000000001Z 1.0 "m1" "f1"
1970-01-01T00:00:00.000000001Z 2.0 "m1" "f2"
1970-01-01T00:00:00.000000001Z null "m1" "f3"
1970-01-01T00:00:00.000000001Z 3.0 "m1" null
1970-01-01T00:00:00.000000002Z 4.0 "m1" "f1"
1970-01-01T00:00:00.000000002Z 5.0 "m1" "f2"
null 6.0 "m1" "f2"
1970-01-01T00:00:00.000000002Z null "m1" "f3"
1970-01-01T00:00:00.000000003Z null "m1" "f1"
1970-01-01T00:00:00.000000003Z 7.0 "m1" null
1970-01-01T00:00:00.000000004Z 8.0 "m1" "f3"

Output:

_time _measurement f1 f2 f3 null
1970-01-01T00:00:00.000000001Z "m1" 1.0 2.0 null 3.0
1970-01-01T00:00:00.000000002Z "m1" 4.0 5.0 null null
null "m1" null 6.0 null null
1970-01-01T00:00:00.000000003Z "m1" null null null 7.0
1970-01-01T00:00:00.000000004Z "m1" null null 8.0 null

Example 2, align fields and measurements that have the same timestamp.
Note the effect of:

  • having null values in some columnKey value;
  • having more values for the same rowKey and columnKey value (the 11th row overrides the 10th, and so does the 15th with the 14th).
  from(bucket:"test")
      |> range(start: 1970-01-01T00:00:00.000000000Z)
      |> pivot(rowKey:["_time"], columnKey: ["_measurement", _field"], valueColumn: "_value")

Input:

_time _value _measurement _field
1970-01-01T00:00:00.000000001Z 1.0 "m1" "f1"
1970-01-01T00:00:00.000000001Z 2.0 "m1" "f2"
1970-01-01T00:00:00.000000001Z 3.0 null "f3"
1970-01-01T00:00:00.000000001Z 4.0 null null
1970-01-01T00:00:00.000000002Z 5.0 "m1" "f1"
1970-01-01T00:00:00.000000002Z 6.0 "m1" "f2"
1970-01-01T00:00:00.000000002Z 7.0 "m1" "f3"
1970-01-01T00:00:00.000000002Z 8.0 null null
null 9.0 "m1" "f3"
1970-01-01T00:00:00.000000003Z 10.0 "m1" null
1970-01-01T00:00:00.000000003Z 11.0 "m1" null
1970-01-01T00:00:00.000000003Z 12.0 "m1" "f3"
1970-01-01T00:00:00.000000003Z 13.0 null null
null 14.0 "m1" null
null 15.0 "m1" null

Output:

_time m1_f1 m1_f2 null_f3 null_null m1_f3 m1_null
1970-01-01T00:00:00.000000001Z 1.0 2.0 3.0 4.0 null null
1970-01-01T00:00:00.000000002Z 5.0 6.0 null 8.0 7.0 null
null null null null null 9.0 15.0
1970-01-01T00:00:00.000000003Z null null null 13.0 12.0 11.0

Join

Join merges two or more input streams, whose values are equal on a set of common columns, into a single output stream. Null values are not considered equal when comparing column values. The resulting schema is the union of the input schemas, and the resulting group key is the union of the input group keys.

Join has the following properties:

Name Type Description
tables object Tables is the map of streams to be joined.
on []string On is the list of columns on which to join.
method string Method must be one of: inner, cross, left, right, or full. Defaults to "inner" .

Both tables and on are required parameters. The on parameter and the cross method are mutually exclusive. Join currently only supports two input streams.

IMPL#83 Add support for joining more than 2 streams
IMPL#84 Add support for different join types

Example:

Given the following two streams of data:

  • SF_Temperature

    _time _field _value
    0001 "temp" 70
    0002 "temp" 75
    0003 "temp" 72
  • NY_Temperature

    _time _field _value
    0001 "temp" 55
    0002 "temp" 56
    0003 "temp" 55

And the following join query:

join(tables: {sf: SF_Temperature, ny: NY_Temperature}, on: ["_time", "_field"])

The output will be:

_time _field _value_ny _value_sf
0001 "temp" 55 70
0002 "temp" 56 75
0003 "temp" 55 72
output schema

The column schema of the output stream is the union of the input schemas, and the same goes for the output group key. Columns that must be renamed due to ambiguity (i.e. columns that occur in more than one input stream) are renamed according to the template <column>_<table>.

Example:

  • SF_Temperature

  • Group Key for table { _field }

    _time _field _value
    0001 "temp" 70
    0002 "temp" 75
    0003 "temp" 72
  • NY_Temperature

  • Group Key for all tables { _time, _field }

    _time _field _value
    0001 "temp" 55
    _time _field _value
    0002 "temp" 56
    _time _field _value
    0003 "temp" 55

join(tables: {sf: SF_Temperature, ny: NY_Temperature}, on: ["_time"]) produces:

  • Group Key for all tables { _time, _field_ny, _field_sf }

    _time _field_ny _field_sf _value_ny _value_sf
    0001 "temp" "temp" 55 70
    _time _field_ny _field_sf _value_ny _value_sf
    0002 "temp" "temp" 56 75
    _time _field_ny _field_sf _value_ny _value_sf
    0003 "temp" "temp" 55 72

Union

Union concatenates two or more input streams into a single output stream. In tables that have identical schema and group keys, contents of the tables will be concatenated in the output stream. The output schemas of the Union operation shall be the union of all input schemas.

Union does not preserve the sort order of the rows within tables. A sort operation may be added if a specific sort order is needed.

Union has the following properties:

Name Type Description
tables []stream Tables specifies the streams to union together. There must be at least two streams.

For example, given this stream, SF_Weather with group key "_field" on both tables:

_time _field _value
0001 "temp" 70
0002 "temp" 75
_time _field _value
0001 "humidity" 81
0002 "humidity" 82

And this stream, NY_Weather, also with group key "_field" on both tables:

_time _field _value
0001 "temp" 55
0002 "temp" 56
_time _field _value
0001 "pressure" 29.82
0002 "pressure" 30.01

union(tables: [SF_Weather, NY_Weather]) produces this stream (whose tables are grouped by "_field"):

_time _field _value
0001 "temp" 70
0002 "temp" 75
0001 "temp" 55
0002 "temp" 56
_time _field _value
0001 "humidity" 81
0002 "humidity" 82
_time _field _value
0001 "pressure" 29.82
0002 "pressure" 30.01

Unique

Unique returns a table with unique values in a specified column. In the case there are multiple rows taking on the same value in the provided column, the first row is kept and the remaining rows are discarded.

Unique has the following properties:

Name Type Description
column string Column that is to have unique values. Defaults to _value.

Cumulative sum

Cumulative sum computes a running sum for non null records in the table. The output table schema will be the same as the input table.

Cumulative sum has the following properties:

Name Type Description
columns []string Columns is a list of columns on which to operate. Defaults to ["_value"].

Example:

from(bucket: "telegraf/autogen")
    |> range(start: -5m)
    |> filter(fn: (r) => r._measurement == "disk" and r._field == "used_percent")
    |> cumulativeSum(columns: ["_value"])

Derivative

Derivative computes the time based difference between subsequent non-null records. This function will return an error if values in the time column are null or not sorted in ascending order. If there are multiple rows with the same time value, only the first row will be used to compute the derivative.

Derivative has the following properties:

Name Type Description
unit duration Unit is the time duration to use for the result. Defaults to 1s.
nonNegative bool NonNegative indicates if the derivative is allowed to be negative. If a value is encountered which is less than the previous value, then the derivative will be null for that row.
columns []string Columns is a list of columns on which to compute the derivative Defaults to ["_value"].
timeColumn string TimeColumn is the column name for the time values. Defaults to _time.
from(bucket: "telegraf/autogen")
    |> range(start: -5m)
    |> filter(fn: (r) => r._measurement == "disk" and r._field == "used_percent")
    |> derivative(nonNegative: true, columns: ["used_percent"])

Difference

Difference computes the difference between subsequent records.
Every user-specified column of numeric type is subtracted while others are kept intact.

Difference has the following properties:

Name Type Description
nonNegative bool NonNegative indicates if the difference is allowed to be negative. If a value is encountered which is less than the previous value then the result is null.
columns []string Columns is a list of columns on which to compute the difference. Defaults to ["_value"].

Rules for subtracting values for numeric types:

  • the difference between two non-null values is their algebraic difference; or null, if the result is negative and nonNegative: true;
  • null minus some value is always null;
  • some value v minus null is v minus the last non-null value seen before v; or null if v is the first non-null value seen.

Example of difference:

_time A B C tag
0001 null 1 2 tv
0002 6 2 null tv
0003 4 2 4 tv
0004 10 10 2 tv
0005 null null 1 tv

Result (nonNegative: false):

_time A B C tag
0002 null 1 null tv
0003 -2 0 2 tv
0004 6 8 -2 tv
0005 null null -1 tv

Result (nonNegative: true):

_time A B C tag
0002 null 1 null tv
0003 null 0 2 tv
0004 6 8 null tv
0005 null null null tv

Example of script:

from(bucket: "telegraf/autogen")
    |> range(start: -5m)
    |> filter(fn: (r) => r._measurement == "cpu" and r._field == "usage_user")
    |> difference()

Increase

Increase returns the total non-negative difference between values in a table. A main usage case is tracking changes in counter values which may wrap over time when they hit a threshold or are reset. In the case of a wrap/reset, we can assume that the absolute delta between two points will be at least their non-negative difference.

increase = (tables=<-, columns=["_value"]) =>
    tables
        |> difference(nonNegative: true, columns:columns)
        |> cumulativeSum()

Example:

Given the following input table.

| _time | _value |
| ----- | ------ |
| 00001 | 1      |
| 00002 | 5      |
| 00003 | 3      |
| 00004 | 4      |

increase() produces the following table.

| _time | _value |
| ----- | ------ |
| 00002 | 4      |
| 00003 | 7      |
| 00004 | 8      |

Distinct

Distinct produces the unique values for a given column. Null is considered its own distinct value if it is present.

Distinct has the following properties:

Name Type Description
column string Column is the column on which to track unique values. Defaults to _value.

Example:

from(bucket: "telegraf/autogen")
	|> range(start: -5m)
	|> filter(fn: (r) => r._measurement == "cpu")
	|> distinct(column: "host")

TimeShift

TimeShift adds a fixed duration to time columns. The output table schema is the same as the input table. If the time is null, the time will continue to be null.

TimeShift has the following properties:

Name Type Description
duration duration Duration is the amount to add to each time value. May be a negative duration.
columns []string Columns is list of all columns that should be shifted. Defaults to ["_start", "_stop", "_time"].

Example:

from(bucket: "telegraf/autogen")
	|> range(start: -5m)
	|> timeShift(duration: 1000h)

StateCount

StateCount computes the number of consecutive records in a given state. The state is defined via a user-defined predicate. For each consecutive point for which the predicate evaluates as true, the state count will be incremented. When a point evaluates as false, the state count is reset.

The state count will be added as an additional column to each record. If the expression evaluates as false, the value will be -1. If the expression generates an error during evaluation, the point is discarded, and does not affect the state count.

StateCount has the following parameters:

Name Type Description
fn (r: record) -> bool Fn is a function that returns true when the record is in the desired state.
column string Column is the name of the column to use to output the state count. Defaults to stateCount.

Example:

from(bucket: "telegraf/autogen")
    |> range(start: 2018-05-22T19:53:26Z)
    |> stateCount(fn:(r) => r._value > 80)

StateDuration

StateDuration computes the duration of a given state. The state is defined via a user-defined predicate. For each consecutive point for which the predicate evaluates as true, the state duration will be incremented by the duration between points. When a point evaluates as false, the state duration is reset.

The state duration will be added as an additional column to each record. If the expression evaluates as false, the value will be -1. If the expression generates an error during evaluation, the point is discarded, and does not affect the state duration.

Note that as the first point in the given state has no previous point, its state duration will be 0.

The duration is represented as an integer in the units specified.

StateDuration requires sorted and not-null timestamps. So, if one of this requirements is not met, it returns an error.

StateDuration has the following parameters:

Name Type Description
fn (r: record) -> bool Fn is a function that returns true when the record is in the desired state.
column string Column is the name of the column to use to output the state value. Defaults to stateDuration.
timeColumn string TimeColumn is the name of the column used to extract timestamps. Defaults to _time.
unit duration Unit is the dimension of the output value. Defaults to 1s.

Example:

from(bucket: "telegraf/autogen")
    |> range(start: 2018-05-22T19:53:26Z)
    |> stateDuration(fn:(r) => r._value > 80)

To

The To operation takes data from a stream and writes it to a bucket. To has the following properties:

Name Type Description
bucket string Bucket is the bucket name into which data will be written.
bucketID string BucketID is the bucket ID into which data will be written.
org string Org is the organization name of the bucket.
orgID string OrgID is the organization ID of the bucket.
host string Host is the location of a remote host to write to. Defaults to "".
token string Token is the authorization token to use when writing to a remote host. Defaults to "".
timeColumn string TimeColumn is the name of the time column of the output. Defaults to "_time".
tagColumns []string TagColumns is a list of columns to be used as tags in the output. Defaults to all columns of type string, excluding all value columns and the _field column if present.
fieldFn (r: record) -> record Function that takes a record from the input table and returns an object. For each record from the input table fieldFn returns on object that maps output field key to output value. Default: (r) => ({ [r._field]: r._value })

TODO(nathanielc): The fieldFn is not valid and needs to change. It uses dynamic object keys which is not allowed.

Either bucket or bucketID is required. Both are mutually exclusive. Similarly org and orgID are mutually exclusive and only required when writing to a remote host. Both host and token are optional parameters, however if host is specified, token is required.

For example, given the following table:

_time _start _stop _measurement _field _value
0005 0000 0009 "a" "temp" 100.1
0006 0000 0009 "a" "temp" 99.3
0007 0000 0009 "a" "temp" 99.9

The default to operation to(bucket:"my-bucket", org:"my-org") is equivalent to writing the above data using the following line protocol:

_measurement=a temp=100.1 0005
_measurement=a temp=99.3 0006
_measurement=a temp=99.9 0007

For an example overriding to's default settings, given the following table:

_time _start _stop tag1 tag2 hum temp
0005 0000 0009 "a" "b" 55.3 100.1
0006 0000 0009 "a" "b" 55.4 99.3
0007 0000 0009 "a" "b" 55.5 99.9

The operation to(bucket:"my-bucket", org:"my-org", tagColumns:["tag1"], fieldFn: (r) => return {"hum": r.hum, "temp": r.temp}) is equivalent to writing the above data using the following line protocol:

_tag1=a hum=55.3,temp=100.1 0005
_tag1=a hum=55.4,temp=99.3 0006
_tag1=a hum=55.5,temp=99.9 0007

Note: The to function produces side effects.

Top/Bottom

Top and Bottom sort a table and limits the table to only n records.

Top and Bottom have the following parameters:

Name Type Description
n int N is the number of records to keep.
columns []string Columns provides the sort order for the tables.

Example:

from(bucket:"telegraf/autogen")
    |> range(start: -5m)
    |> filter(fn:(r) => r._measurement == "net" and r._field == "bytes_sent")
    |> top(n:10, columns:["_value"])

Contains

Tests whether a value is a member of a set.

Contains has the following parameters:

Name Type Description
value bool, int, uint, float, string, time The value to search for.
set array of bool, int, uint, float, string, time The set of values to search.

Example: contains(value:1, set:[1,2,3]) will return true.

Stream/table functions

These functions allow to extract a table from a stream of tables (tableFind) and access its columns (getColumn) and records (getRecord).
The example below provides an overview of these functions, further information can be found in the paragraphs below.

data = from(bucket:"telegraf/autogen")
    |> range(start: -5m)
    |> filter(fn:(r) => r._measurement == "cpu")

// extract the first available table for which "_field" is equal to "usage_idle"
t = data |> tableFind(fn: (key) => key._field == "usage_idle")
// this is the same as 'tableFind(tables: data, fn: (key) => key._field == "usage_idle")'

// extract the "_value" column from the table
values = t |> getColumn(column: "_value")
// the same as 'getColumn(table: t, column: "_value")'

// extract the first record from the table
r0 = t |> getRecord(idx: 0)
// the same as 'getRecord(table: t, idx: 0)'
TableFind

TableFind extracts the first table in a stream of table whose group key values match a given predicate. If no table is found, the function errors.

It has the following parameters:

Name Type Description
fn (key: object) -> bool Fn is a predicate function. The result is the first table for which fn is true.

Example:

t = from(bucket:"telegraf/autogen")
    |> range(start: -5m)
    |> filter(fn:(r) => r._measurement == "cpu")
    |> tableFind(fn: (key) => key._field == "usage_idle")

// `t` contains the first table whose group key value for `_field` is "usage_idle" after filtering.
// `t` can be used as input to `getColumn` and `getRecord`.
GetColumn

GetColumn extracts a column from a table given its label. If the label is not present in the set of columns, the function errors.

It has the following parameters:

Name Type Description
column string Column is the name of the column to extract.

Example:

vs = from(bucket:"telegraf/autogen")
    |> range(start: -5m)
    |> filter(fn:(r) => r._measurement == "cpu")
    |> tableFind(fn: (key) => key._field == "usage_idle")
    |> getColumn(column: "_value")

// use values
x = vs[0] + vs[1]
GetRecord

GetRecord extracts a record from a table given its index. If the index is out of bounds, the function errors.

It has the following parameters:

Name Type Description
idx int Idx is the index of the record to extract.

Example:

r0 = from(bucket:"telegraf/autogen")
    |> range(start: -5m)
    |> filter(fn:(r) => r._measurement == "cpu")
    |> tableFind(fn: (key) => key._field == "usage_idle")
    |> getRecord(idx: 0)

// use values
x = r0._field + "--" + r0._measurement

Type conversion operations

toBool

Convert a value to a bool.

Example: from(bucket: "telegraf") |> filter(fn:(r) => r._measurement == "mem" and r._field == "used") |> toBool()

The function toBool is defined as toBool = (tables=<-) => tables |> map(fn:(r) => bool(v:r._value)). If you need to convert other columns use the map function directly with the bool function.

toInt

Convert a value to a int.

Example: from(bucket: "telegraf") |> filter(fn:(r) => r._measurement == "mem" and r._field == "used") |> toInt()

The function toInt is defined as toInt = (tables=<-) => tables |> map(fn:(r) => int(v:r._value)). If you need to convert other columns use the map function directly with the int function.

toFloat

Convert a value to a float.

Example: from(bucket: "telegraf") |> filter(fn:(r) => r._measurement == "mem" and r._field == "used") |> toFloat()

The function toFloat is defined as toFloat = (tables=<-) => tables |> map(fn:(r) => float(v:r._value)). If you need to convert other columns use the map function directly with the float function.

toDuration

Convert a value to a duration.

Example: from(bucket: "telegraf") |> filter(fn:(r) => r._measurement == "mem" and r._field == "used") |> toDuration()

The function toDuration is defined as toDuration = (tables=<-) => tables |> map(fn:(r) => duration(v:r._value)). If you need to convert other columns use the map function directly with the duration function.

toString

Convert a value to a string.

Example: from(bucket: "telegraf") |> filter(fn:(r) => r._measurement == "mem" and r._field == "used") |> toString()

The function toString is defined as toString = (tables=<-) => tables |> map(fn:(r) => string(v:r._value)). If you need to convert other columns use the map function directly with the string function.

toTime

Convert a value to a time.

Example: from(bucket: "telegraf") |> filter(fn:(r) => r._measurement == "mem" and r._field == "used") |> toTime()

The function toTime is defined as toTime = (tables=<-) => tables |> map(fn:(r) => time(v:r._value)). If you need to convert other columns use the map function directly with the time function.

toUInt

Convert a value to a uint.

Example: from(bucket: "telegraf") |> filter(fn:(r) => r._measurement == "mem" and r._field == "used") |> toUInt()

The function toUInt is defined as toUInt = (tables=<-) => tables |> map(fn:(r) => uint(v:r._value)). If you need to convert other columns use the map function directly with the uint function.

IMPL#242 Update specification around type conversion functions.

String operations

compare

Compare two strings lexicographically. Returns 0 if v==t, -1 if v < t, and +1 if v > t.

Example: compare(v: "a", t: "a") returns the int 0

containsStr

Reports whether substr is in v.

Example: containsStr(v: "abc", substr: "a") returns the boolean true.

containsAny

Reports whether any value from chars is in v.

Example: containsAny(v: "abc", chars: "and") returns the boolean true.

countStr

Reports the number of non-overlapping instances in which substr appears in v.

Example: countStr(v: "aaaaa", substr: "a") returns the int 5.

equalFold

Reports whether v and t, interpreted as UTF-8 strings, are equal under Unicode case-folding.

Example: equalFold(v: "Go", t: "go") returns boolean true.

hasPrefix

Tests whether the string v begins with prefix.

Example: hasPrefix(v: "go gopher", t: "go") returns boolean true.

hasSuffix

Tests whether the string v ends with suffix.

Example: hasSuffix(v: "go gopher", t: "go") returns boolean false.

index

Returns the index of the first instance of substr in v, or -1 if substr is not present in s.

Example: index(v: "go gopher", substr: "go") returns int 0.

indexAny

Returns the index of the first instance of any value in substr in v, or -1 if all values in substr are not present in s.

Example: indexAny(v: "chicken", chars: "aeiouy") returns int 2.

isDigit

Returns whether or not v is a digit.

Example: isDigit(v: "A") returns boolean false.

isLetter

Returns whether or not v is a letter.

Example: isLetter(v: "A") returns boolean true.

isLower

Returns whether or not v is lowercase.

Example: isLower(v: "A") returns boolean false.

isUpper

Returns whether or not v is uppercase.

Example: isUpper(v: "A") returns boolean true.

joinStr

Concatenates the elements of inputted string array to create a single string with v as the separator.

Example: joinStr(arr: []string{"a", "b", "c"}, v: ",") returns string a,b,c.

lastIndex

Returns the index of the last instance of substr in s, or -1 if substr is not present in v.

Example: lastIndex(v: "go gopher", t: "go") returns int 3.

lastIndexAny

Returns the index of the last instance of any value from chars in v, or -1 if no value from chars is present in v.

Example: lastIndexAny(v: "go gopher", t: "go") returns int 4.

repeat

Returns a new string consisting of i copies of the string v.

Example: repeat("v: na", i: 2) returns string nana.

replace

Returns a copy of the string v with the first i non-overlapping instances of t replaced by u.

Example: replaceAll(v: "oink oink oink", t: "oink", u: "moo", i: 2) returns string moo moo oink.

replaceAll

Returns a copy of the string v with the all non-overlapping instances of t replaced by u.

Example: replaceAll(v: "oink oink oink", t: "oink", u: "moo") returns string moo moo moo.

split

Slices v into all substrings separated by t and returns a slice of the substrings between those separators.

Example: split(v: "a,b,c", t: ",") returns []string ["a" "b" "c"].

splitAfter

Slices v into all substrings after each instance of t and returns a slice of the substrings between those separators.

Example: splitAfter(v: "a,b,c", t: ",") returns []string ["a," "b," "c"].

splitAfterN

Slices v into all substrings after each instance of t and returns a slice of the substrings between those separators. i determines the number of substrings to return.

Example: splitAfterN(v: "a,b,c", t: ",", i: 2) returns []string ["a," "b,c"].

splitN

Slices v into all substrings separated by t and returns a slice of the substrings between those separators. i determines the number of substrings to return.

Example: splitN(v: "a,b,c", t: ",", i: 2) returns []string ["a" "b,c"].

title

Convert a string to title case.

Example: title(v: "a flux of foxes") returns the string A Flux Of Foxes.

toLower

Convert a string to lower case.

Example: toLower(v: "KOALA") returns the string koala.

toTitle

Returns a copy of the string v with all Unicode letters mapped to their title case.

Example: toTitle("loud noises") returns the string LOUD NOISES.

toUpper

Convert a string to upper case.

Example: toUpper(v: "koala") returns the string KOALA.

trim

Remove leading and trailing characters specified in cutset from a string.

Example: trim(v: ".abc.", cutset: ".") returns the string abc.

trimLeft

Remove leading characters specified in cutset from a string.

Example: trim(v: ".abc.", cutset: ".") returns the string abc..

trimPrefix

Remove a prefix from a string. Strings that do not start with the prefix are returned unchanged.

Example: trimPrefix(v: "123_abc", prefix: "123") returns the string _abc.

trimRight

Remove trailing characters specified in cutset from a string.

Example: trim(v: ".abc.", cutset: ".") returns the string .abc.

trimSpace

Remove leading and trailing spaces from a string.

Example: trimSpace(v: " abc ") returns the string abc.

trimSuffix

Remove a suffix from a string. Strings that do not end with the suffix are returned unchanged.

Example: trimSuffix(v: "abc_123", suffix: "123") returns the string abc_.

Regexp Operations

compile

Parse a regular expression and return, if successful, a Regexp object that can be used to match against text.

Example: compile(v: "abcd") returns the Regex object abcd.

findString

Return a string holding the text of the leftmost match in v of the regular expression.

Example: findString(r: regexp.compile("foo.?"), v: "seafood fool") returns the string food.

findStringIndex

Returns a two-element slice of integers defining the location of the leftmost match in v of the regular expression.

Example: findStringIndex(r: regexp.compile("ab?"), v: "tablett") returns the int array [1 3].

getString

Return the source text used to compile the regular expression.

Example: getString(v: regexp.compile("abcd")) returns the Regex object abcd.

matchRegexString

Report whether the string v contains any match of the regular expression r.

Example: matchString(r: regexp.compile("(gopher){2}"), v: "gophergophergopher") returns boolean true

replaceAllString

Returns a copy of v, replacing matches of the Regexp r with the replacement string t.

Example: replaceAllString(r: regexp.compile("a(x*)b"), v: "-ab-axxb-", t: "T") returns string -T-T-.

quoteMeta

Return a string that escapes all regular expression metacharacters inside the argument text; the returned string is a regular expression matching the literal text.

Example: quoteMeta("Escaping symbols like: .+*?()|[]{}^$") returns string Escaping symbols like: \.\+\*\?\(\)\|\[\]\{\}\^\$.

splitRegexp

Slices v into substrings separated by the expression and returns a slice of the substrings between those expression matches.

Example: splitRegex(r: regexp.compile("a*"), v: "abaabaccadaaae", i: 5) returns string array ["", "b", "b", "c", "cadaaae"].

Composite data types

A composite data type is a collection of primitive data types that together have a higher meaning.

Triggers

A trigger is associated with a table and contains logic for when it should fire. When a trigger fires its table is materialized. Materializing a table makes it available for any down stream operations to consume. Once a table is materialized it can no longer be modified.

Triggers can fire based on these inputs:

Input Description
Current processing time The current processing time is the system time when the trigger is being evaluated.
Watermark time The watermark time is a time where it is expected that no data will arrive that is older than it.
Record count The number of records currently in the table.
Group key value The group key value of the table.

Additionally triggers can be finished, which means that they will never fire again. Once a trigger is finished, its associated table is deleted.

Currently all tables use an after watermark trigger which fires only once the watermark has exceeded the _stop value of the table and then is immediately finished.

Data sources are responsible for informing about updates to the watermark.

IMPL#240 Make trigger support not dependent on specific columns

Execution model

A query specification defines what data and operations to perform. The execution model reserves the right to perform those operations as efficiently as possible. The execution model may rewrite the query in anyway it sees fit while maintaining correctness.

Request and Response Formats

Included with the specification of the language and execution model, is a specification of how to submit queries and read their responses over HTTP.

Request format

To submit a query for execution, make an HTTP POST request to the /v1/query endpoint.

The POST request may either submit parameters as the POST body or a subset of the parameters as URL query parameters. The following parameters are supported:

Parameter Description
query Query is Flux text describing the query to run. Only one of query or spec may be specified. This parameter may be passed as a URL parameter.
spec Spec is a query specification. Only one of query or spec may be specified.
dialect Dialect is an object defining the options to use when encoding the response.

When using the POST body to submit the query the Content-Type HTTP header must contain the name of the request encoding being used.

Supported request content types:

  • application/json - Use a JSON encoding of parameters.

Multiple response content types will be supported. The desired response content type is specified using the Accept HTTP header on the request. Each response content type will have its own dialect options.

Supported response encodings:

  • test/csv - Corresponds with the MIME type specified in RFC 4180. Details on the encoding format are specified below.

If no Accept header is present it is assumed that text/csv was specified. The HTTP header Content-Type of the response will specify the encoding of the response.

Examples requests

Make a request using a query string and URL query parameters:

POST /v1/query?query=%20from%28db%3A%22mydatabse%22%29%20%7C%3E%20last%28%29 HTTP/1.1

Make a request using a query string and the POST body as JSON:

POST /v1/query


{
    "query": "from(bucket:\"mydatabase/autogen\") |> last()"
}

Make a request using a query specification and the POST body as JSON:

POST /v1/query


{
    "spec": {
      "operations": [
        {
          "kind": "from",
          "id": "from0",
          "spec": {
            "db": "mydatabase"
          }
        },
        {
          "kind": "last",
          "id": "last1",
          "spec": {
            "column": ""
          }
        }
      ],
      "edges": [
        {
          "parent": "from0",
          "child": "last1"
        }
      ],
      "resources": {
        "priority": "high",
        "concurrency_quota": 0,
        "memory_bytes_quota": 0
      }
    }
}

Make a request using a query string and the POST body as JSON. Dialect options are specified for the text/csv format. See below for details on specific dialect options.

POST /v1/query


{
    "query": "from(bucket:\"mydatabase/autogen\") |> last()",
    "dialect" : {
        "header": true,
        "annotations": ["datatype"]
    }
}

Response format

CSV

The result of a query is any number of named streams. As a stream consists of multiple tables each table is encoded as CSV textual data. CSV data should be encoded using UTF-8, and should be in Unicode Normal Form C as defined in UAX15. Line endings must be CRLF as defined by the text/csv MIME type in RFC 4180

Each table may have the following rows:

  • annotation rows - a set of rows describing properties about the columns of the table.
  • header row - a single row that defines the column labels.
  • record rows, a set of rows containing the record data, one record per row.

In addition to the columns on the tables themselves three additional columns may be added to the CSV table.

  • annotation - Contains the name of an annotation. This column is optional, if it exists it is always the first column. The only valid values for the column are the list of supported annotations or an empty value.
  • result - Contains the name of the result as specified by the query.
  • table - Contains a unique ID for each table within a result.

Columns support the following annotations:

  • datatype - a description of the type of data contained within the column.
  • group - a boolean flag indicating if the column is part of the table's group key.
  • default - a default value to be used for rows whose string value is the empty string.
  • null - a value indicating that data is missing.
Multiple tables

Multiple tables may be encoded into the same file or data stream. The table column indicates the table a row belongs to. All rows for a table must be contiguous.

It is possible that multiple tables in the same result do not share a common table scheme. It is also possible that a table has no records. In such cases an empty row delimits a new table boundary and new annotations and header rows follow. The empty row acts like a delimiter between two independent CSV files that have been concatenated together.

In the case were a table has no rows the default annotation is used to provide the values of the group key.

Multiple results

Multiple results may be encoded into the same file or data stream. An empty row always delimits separate results within the same file. The empty row acts like a delimiter between two independent CSV files that have been concatenated together.

Annotations

Annotations rows are prefixed with a comment marker. The first column contains the name of the annotation being defined. The subsequent columns contain the value of the annotation for the respective columns.

The datatype annotation specifies the data types of the remaining columns. The possible data types are:

Datatype Flux type Description
boolean bool a truth value, one of "true" or "false"
unsignedLong uint an unsigned 64-bit integer
long int a signed 64-bit integer
double float a IEEE-754 64-bit floating-point number
string string a UTF-8 encoded string
base64Binary bytes a base64 encoded sequence of bytes as defined in RFC 4648
dateTime time an instant in time, may be followed with a colon : and a description of the format
duration duration a length of time represented as an unsigned 64-bit integer number of nanoseconds

The group annotation specifies if the column is part of the table's group key. Possible values are true or false.

The default annotation specifies a default value, if it exists, for each column.

In order to fully encode a table with its group key the datatype, group and default annotations must be used.

The null annotation specifies the string value that indicates a missing value. When the null annotation is not specified the empty string value is the null value for the column. It is not possible to encode/decode a non-null string value that is the same as the null annotation value for columns of type string.

When the default annotation value of a column is the same as the null annotation value of a column, it is interpreted as the column's default value is null.

Errors

When an error occurs during execution a table will be returned with the first column label as error and the second column label as reference. The error's properties are contained in the second row of the table. The error column contains the error message and the reference column contains a unique reference code that can be used to get further information about the problem.

When an error occurs before any results are materialized then the HTTP status code will indicate an error and the error details will be encoded in the csv table. When an error occurs after some results have already been sent to the client the error will be encoded as the next table and the rest of the results will be discarded. In such a case the HTTP status code cannot be changed and will remain as 200 OK.

Example error encoding without annotations:

error,reference
Failed to parse query,897
Dialect options

The CSV response format support the following dialect options:

Option Description
header Header is a boolean value, if true the header row is included, otherwise its is omitted. Defaults to true.
delimiter Delimiter is a character to use as the delimiting value between columns. Defaults to ",".
quoteChar QuoteChar is a character to use to quote values containing the delimiter. Defaults to ".
annotations Annotations is a list of annotations that should be encoded. If the list is empty the annotation column is omitted entirely. Defaults to an empty list.
commentPrefix CommentPrefix is a string prefix to add to comment rows. Defaults to "#". Annotations are always comment rows.
Examples

For context the following example tables encode fictitious data in response to this query:

from(bucket:"mydb/autogen")
    |> range(start:2018-05-08T20:50:00Z, stop:2018-05-08T20:51:00Z)
    |> group(columns:["_start","_stop", "region", "host"])
    |> mean()
    |> group(columns:["_start","_stop", "region"])
    |> yield(name:"mean")

Example encoding with of a single table with no annotations:

result,table,_start,_stop,_time,region,host,_value
mean,0,2018-05-08T20:50:00Z,2018-05-08T20:51:00Z,2018-05-08T20:50:00Z,east,A,15.43
mean,0,2018-05-08T20:50:00Z,2018-05-08T20:51:00Z,2018-05-08T20:50:20Z,east,B,59.25
mean,0,2018-05-08T20:50:00Z,2018-05-08T20:51:00Z,2018-05-08T20:50:40Z,east,C,52.62

Example encoding with two tables in the same result with no annotations:

result,table,_start,_stop,_time,region,host,_value
mean,0,2018-05-08T20:50:00Z,2018-05-08T20:51:00Z,2018-05-08T20:50:00Z,east,A,15.43
mean,0,2018-05-08T20:50:00Z,2018-05-08T20:51:00Z,2018-05-08T20:50:20Z,east,B,59.25
mean,0,2018-05-08T20:50:00Z,2018-05-08T20:51:00Z,2018-05-08T20:50:40Z,east,C,52.62
mean,1,2018-05-08T20:50:00Z,2018-05-08T20:51:00Z,2018-05-08T20:50:00Z,west,A,62.73
mean,1,2018-05-08T20:50:00Z,2018-05-08T20:51:00Z,2018-05-08T20:50:20Z,west,B,12.83
mean,1,2018-05-08T20:50:00Z,2018-05-08T20:51:00Z,2018-05-08T20:50:40Z,west,C,51.62

Example encoding with two tables in the same result with no annotations and no header row:

mean,0,2018-05-08T20:50:00Z,2018-05-08T20:51:00Z,2018-05-08T20:50:00Z,east,A,15.43
mean,0,2018-05-08T20:50:00Z,2018-05-08T20:51:00Z,2018-05-08T20:50:20Z,east,B,59.25
mean,0,2018-05-08T20:50:00Z,2018-05-08T20:51:00Z,2018-05-08T20:50:40Z,east,C,52.62
mean,1,2018-05-08T20:50:00Z,2018-05-08T20:51:00Z,2018-05-08T20:50:00Z,west,A,62.73
mean,1,2018-05-08T20:50:00Z,2018-05-08T20:51:00Z,2018-05-08T20:50:20Z,west,B,12.83
mean,1,2018-05-08T20:50:00Z,2018-05-08T20:51:00Z,2018-05-08T20:50:40Z,west,C,51.62

Example encoding with two tables in the same result with the datatype annotation:

#datatype,string,long,dateTime:RFC3339,dateTime:RFC3339,dateTime:RFC3339,string,string,double
,result,table,_start,_stop,_time,region,host,_value
,mean,0,2018-05-08T20:50:00Z,2018-05-08T20:51:00Z,2018-05-08T20:50:00Z,east,A,15.43
,mean,0,2018-05-08T20:50:00Z,2018-05-08T20:51:00Z,2018-05-08T20:50:20Z,east,B,59.25
,mean,0,2018-05-08T20:50:00Z,2018-05-08T20:51:00Z,2018-05-08T20:50:40Z,east,C,52.62
,mean,1,2018-05-08T20:50:00Z,2018-05-08T20:51:00Z,2018-05-08T20:50:00Z,west,A,62.73
,mean,1,2018-05-08T20:50:00Z,2018-05-08T20:51:00Z,2018-05-08T20:50:20Z,west,B,12.83
,mean,1,2018-05-08T20:50:00Z,2018-05-08T20:51:00Z,2018-05-08T20:50:40Z,west,C,51.62

Example encoding with two tables in the same result with the datatype and group annotations:

#datatype,string,long,dateTime:RFC3339,dateTime:RFC3339,dateTime:RFC3339,string,string,double
#group,false,false,true,true,false,true,false,false
,result,table,_start,_stop,_time,region,host,_value
,mean,0,2018-05-08T20:50:00Z,2018-05-08T20:51:00Z,2018-05-08T20:50:00Z,east,A,15.43
,mean,0,2018-05-08T20:50:00Z,2018-05-08T20:51:00Z,2018-05-08T20:50:20Z,east,B,59.25
,mean,0,2018-05-08T20:50:00Z,2018-05-08T20:51:00Z,2018-05-08T20:50:40Z,east,C,52.62
,mean,1,2018-05-08T20:50:00Z,2018-05-08T20:51:00Z,2018-05-08T20:50:00Z,west,A,62.73
,mean,1,2018-05-08T20:50:00Z,2018-05-08T20:51:00Z,2018-05-08T20:50:20Z,west,B,12.83
,mean,1,2018-05-08T20:50:00Z,2018-05-08T20:51:00Z,2018-05-08T20:50:40Z,west,C,51.62

Example encoding with two tables with differing schemas in the same result with the datatype and group annotations:

#datatype,string,long,dateTime:RFC3339,dateTime:RFC3339,dateTime:RFC3339,string,string,double
#group,false,false,true,true,false,true,false,false
,result,table,_start,_stop,_time,region,host,_value
,mean,0,2018-05-08T20:50:00Z,2018-05-08T20:51:00Z,2018-05-08T20:50:00Z,east,A,15.43
,mean,0,2018-05-08T20:50:00Z,2018-05-08T20:51:00Z,2018-05-08T20:50:20Z,east,B,59.25
,mean,0,2018-05-08T20:50:00Z,2018-05-08T20:51:00Z,2018-05-08T20:50:40Z,east,C,52.62

#datatype,string,long,dateTime:RFC3339,dateTime:RFC3339,dateTime:RFC3339,string,string,double
#group,false,false,true,true,false,true,false,false
,result,table,_start,_stop,_time,location,device,min,max
,mean,1,2018-05-08T20:50:00Z,2018-05-08T20:51:00Z,2018-05-08T20:50:00Z,USA,5825,62.73,68.42
,mean,1,2018-05-08T20:50:00Z,2018-05-08T20:51:00Z,2018-05-08T20:50:20Z,USA,2175,12.83,56.12
,mean,1,2018-05-08T20:50:00Z,2018-05-08T20:51:00Z,2018-05-08T20:50:40Z,USA,6913,51.62,54.25

Example error encoding with the datatype annotation:

#datatype,string,long
,error,reference
,Failed to parse query,897

Example error encoding with after a valid table has already been encoded.

#datatype,string,long,dateTime:RFC3339,dateTime:RFC3339,dateTime:RFC3339,string,string,double
,result,table,_start,_stop,_time,region,host,_value
,mean,1,2018-05-08T20:50:00Z,2018-05-08T20:51:00Z,2018-05-08T20:50:00Z,west,A,62.73
,mean,1,2018-05-08T20:50:00Z,2018-05-08T20:51:00Z,2018-05-08T20:50:20Z,west,B,12.83
,mean,1,2018-05-08T20:50:00Z,2018-05-08T20:51:00Z,2018-05-08T20:50:40Z,west,C,51.62

#datatype,string,long
,error,reference
,query terminated: reached maximum allowed memory limits,576
You can’t perform that action at this time.