Skip to content
Switch branches/tags


Failed to load latest commit information.
Latest commit message
Commit time
Sep 9, 2021
Sep 9, 2021
Sep 10, 2021


go-mysql-server is a SQL engine which parses standard SQL (based on MySQL syntax) and executes queries on data sources of your choice. A simple in-memory database and table implementation are provided, and you can query any data source you want by implementing a few interfaces.

go-mysql-server also provides a server implementation compatible with the MySQL wire protocol. That means it is compatible with MySQL ODBC, JDBC, or the default MySQL client shell interface.

Dolt, a SQL database with Git-style versioning, is the main database implementation of this package. Check out that project for reference implementations.

Scope of this project

These are the goals of go-mysql-server:

  • Be a generic extensible SQL engine that performs queries on your data sources.
  • Provide a simple database implementation suitable for use in tests.
  • Define interfaces you can implement to query your own data sources.
  • Provide a runnable server speaking the MySQL wire protocol, connected to data sources of your choice.
  • Optimize query plans.
  • Allow implementors to add their own analysis steps and optimizations.
  • Support indexed lookups and joins on data tables that support them.
  • Support external index driver implementations such as pilosa.
  • With few caveats and using a full database implementation, be a drop-in MySQL database replacement.

Non-goals of go-mysql-server:

  • Be an application/server you can use directly.
  • Provide any kind of backend implementation (other than the memory one used for testing) such as json, csv, yaml. That's for clients to implement and use.

What's the use case of go-mysql-server?

go-mysql-server has two primary uses case:

  1. Stand-in for MySQL in a golang test environment, using the built-in memory database implementation.

  2. Providing access to aribtrary data sources with SQL queries by implementing a handful of interfaces. The most complete real-world implementation is Dolt.


The import path for the package is

To install it, run:

go get

Go Documentation

SQL syntax

The goal of go-mysql-server is to support 100% of the statements that MySQL does. We are continuously adding more functionality to the engine, but not everything is supported yet. To see what is currently included check the SUPPORTED file.

Third-party clients

We support and actively test against certain third-party clients to ensure compatibility between them and go-mysql-server. You can check out the list of supported third party clients in the SUPPORTED_CLIENTS file along with some examples on how to connect to go-mysql-server using them.

Available functions

Name Description
ABS(expr) returns the absolute value of an expression
ACOS(expr) returns the arccos of an expression
ARRAY_LENGTH(json) if the json representation is an array, this function returns its size.
ASIN(expr) returns the arcsin of an expression
ATAN(expr) returs the arctan of an expression
AVG(expr) returns the average value of expr in all rows.
CEIL(number) returns the smallest integer value that is greater than or equal to number.
CEILING(number) returns the smallest integer value that is greater than or equal to number.
CHARACTER_LENGTH(str) returns the length of the string in characters.
CHAR_LENGTH(str) returns the length of the string in characters.
COALESCE(...) returns the first non-null value in a list.
CONCAT(...) concatenates any group of fields into a single string.
CONCAT_WS(sep, ...) concatenates any group of fields into a single string. The first argument is the separator for the rest of the arguments. The separator is added between the strings to be concatenated. The separator can be a string, as can the rest of the arguments. If the separator is NULL, the result is NULL.
CONNECTION_ID() returns the current connection ID.
COS(expr) returns the cosine of an expression.
COT(expr) returns the arctangent of an expression.
COUNT(expr) returns a count of the number of non-NULL values of expr in the rows retrieved by a SELECT statement.
CURRENT_USER() returns the current user
DATE(date) returns the date part of the given date.
DATETIME(expr) returns a DATETIME value for the expression given (e.g. the string '2020-01-02').
DATE_ADD(date, interval) adds the interval to the given date.
DATE_SUB(date, interval) subtracts the interval from the given date.
DAY(date) is a synonym for DAYOFMONTH().
DAYOFMONTH(date) returns the day of the month (0-31).
DAYOFWEEK(date) returns the day of the week of the given date.
DAYOFYEAR(date) returns the day of the year of the given date.
DEGREES(expr) returns the number of degrees in the radian expression given.
EXPLODE(...) generates a new row in the result set for each element in the expressions provided.
FIRST(expr) returns the first value in a sequence of elements of an aggregation.
FLOOR(number) returns the largest integer value that is less than or equal to number.
FROM_BASE64(str) decodes the base64-encoded string str.
GREATEST(...) returns the greatest numeric or string value.
HOUR(date) returns the hours of the given date.
IFNULL(expr1, expr2) if expr1 is not NULL, it returns expr1; otherwise it returns expr2.
IF(expr1, expr2, expr3) if expr1 evaluates to true, retuns expr2. Otherwise returns expr3.
INSTR(str1, str2) returns the 1-based index of the first occurence of str2 in str1, or 0 if it does not occur.
IS_BINARY(blob) returns whether a blob is a binary file or not.
ISNULL(expr) returns whether a expr is null or not.
JSON_EXTRACT(json_doc, path, ...) extracts data from a json document using json paths. Extracting a string will result in that string being quoted. To avoid this, use JSON_UNQUOTE(JSON_EXTRACT(json_doc, path, ...)).
JSON_UNQUOTE(json) unquotes JSON value and returns the result as a utf8mb4 string.
LAST(expr) returns the last value in a sequence of elements of an aggregation.
LEAST(...) returns the smaller numeric or string value.
LEFT(str, int) returns the first N characters in the string given.
LENGTH(str) returns the length of the string in bytes.
LN(X) returns the natural logarithm of X.
LOG(X), LOG(B, X) if called with one parameter, this function returns the natural logarithm of X. If called with two parameters, this function returns the logarithm of X to the base B. If X is less than or equal to 0, or if B is less than or equal to 1, then NULL is returned.
LOG10(X) returns the base-10 logarithm of X.
LOG2(X) returns the base-2 logarithm of X.
LOWER(str) returns the string str with all characters in lower case.
LPAD(str, len, padstr) returns the string str, left-padded with the string padstr to a length of len characters.
LTRIM(str) returns the string str with leading space characters removed.
MAX(expr) returns the maximum value of expr in all rows.
MID(str, pos, [len]) returns a substring from the provided string starting at pos with a length of len characters. If no len is provided, all characters from pos until the end will be taken.
MIN(expr) returns the minimum value of expr in all rows.
MINUTE(date) returns the minutes of the given date.
MONTH(date) returns the month of the given date.
NOW() returns the current timestamp.
NULLIF(expr1, expr2) returns NULL if expr1 = expr2 is true, otherwise returns expr1.
POW(X, Y) returns the value of X raised to the power of Y.
POWER(X, Y) synonym for POW
RADIANS(expr) returns the radian value of the degrees argument given
RAND(expr?) returns a random number in the range 0 <= x < 1. If an argument is given, it is used to seed the random number generator.
REGEXP_MATCHES(text, pattern, [flags]) returns an array with the matches of the pattern in the given text. Flags can be given to control certain behaviours of the regular expression. Currently, only the i flag is supported, to make the comparison case insensitive.
REPEAT(str, count) returns a string consisting of the string str repeated count times.
REPLACE(str,from_str,to_str) returns the string str with all occurrences of the string from_str replaced by the string to_str.
REVERSE(str) returns the string str with the order of the characters reversed.
ROUND(number, decimals) rounds the number to decimals decimal places.
RPAD(str, len, padstr) returns the string str, right-padded with the string padstr to a length of len characters.
RTRIM(str) returns the string str with trailing space characters removed.
SECOND(date) returns the seconds of the given date.
SIN(expr) returns the sine of the expression given.
SLEEP(seconds) waits for the specified number of seconds (can be fractional).
SOUNDEX(str) returns the soundex of a string.
SPLIT(str,sep) returns the parts of the string str split by the separator sep as a JSON array of strings.
SQRT(X) returns the square root of a nonnegative number X.
STR_TO_DATE(date_str, format_str) parses the date/datetime/timestamp expression according to the format specifier.
SUBSTR(str, pos, [len]) returns a substring from the string str starting at pos with a length of len characters. If no len is provided, all characters from pos until the end will be taken.
SUBSTRING(str, pos, [len]) returns a substring from the string str starting at pos with a length of len characters. If no len is provided, all characters from pos until the end will be taken.
SUBSTRING_INDEX(str, delim, count) Returns a substring after count appearances of delim. If count is negative, counts from the right side of the string.
SUM(expr) returns the sum of expr in all rows.
TAN(expr) returns the tangent of the expression given.
TIMEDIFF(expr1, expr2) returns expr1 − expr2 expressed as a time value. expr1 and expr2 are time or date-and-time expressions, but both must be of the same type.
TIMESTAMP(expr) returns a timestamp value for the expression given (e.g. the string '2020-01-02').
TO_BASE64(str) encodes the string str in base64 format.
TRIM(str) returns the string str with all spaces removed.
UNIX_TIMESTAMP(expr?) returns the datetime argument to the number of seconds since the Unix epoch. With nor argument, returns the number of execonds since the Unix epoch for the current time.
UPPER(str) returns the string str with all characters in upper case.
USER() returns the current user name.
UTC_TIMESTAMP() returns the current UTC timestamp.
WEEKDAY(date) returns the weekday of the given date.
YEAR(date) returns the year of the given date.
YEARWEEK(date, mode) returns year and week for a date. The year in the result may be different from the year in the date argument for the first and the last week of the year.


The behaviour of certain parts of go-mysql-server can be configured using either environment variables or session variables.

Session variables are set using the following SQL queries:

SET <variable name> = <value>
Name Type Description
INMEMORY_JOINS environment If set it will perform all joins in memory. Default is off.
inmemory_joins session If set it will perform all joins in memory. Default is off. This has precedence over INMEMORY_JOINS.
MAX_MEMORY environment The maximum number of memory, in megabytes, that can be consumed by go-mysql-server. Any in-memory caches or computations will no longer try to use memory when the limit is reached. Note that this may cause certain queries to fail if there is not enough memory available, such as queries using DISTINCT, ORDER BY or GROUP BY with groupings.
DEBUG_ANALYZER environment If set, the analyzer will print debug messages. Default is off.


go-mysql-server contains a SQL engine and server implementation. So, if you want to start a server, first instantiate the engine and pass your sql.Database implementation.

It will be in charge of handling all the logic to retrieve the data from your source. Here you can see an example using the in-memory database implementation:

package main

import (

    sqle ""

func main() {
    driver := sqle.NewDefault()

    config := server.Config{
        Protocol: "tcp",
        Address:  "localhost:3306",
        Auth:     auth.NewNativeSingle("user", "pass", auth.AllPermissions),

    s, err := server.NewDefaultServer(config, driver)
    if err != nil {


func createTestDatabase() *memory.Database {
    const (
        dbName    = "test"
        tableName = "mytable"

    db := memory.NewDatabase(dbName)
    table := memory.NewTable(tableName, sql.Schema{
        {Name: "name", Type: sql.Text, Nullable: false, Source: tableName},
        {Name: "email", Type: sql.Text, Nullable: false, Source: tableName},
        {Name: "phone_numbers", Type: sql.JSON, Nullable: false, Source: tableName},
        {Name: "created_at", Type: sql.Timestamp, Nullable: false, Source: tableName},

    db.AddTable(tableName, table)
    ctx := sql.NewEmptyContext()

    rows := []sql.Row{
        sql.NewRow("John Doe", "", []string{"555-555-555"}, time.Now()),
        sql.NewRow("John Doe", "", []string{}, time.Now()),
        sql.NewRow("Jane Doe", "", []string{}, time.Now()),
        sql.NewRow("Evil Bob", "", []string{"555-666-555", "666-666-666"}, time.Now()),

    for _, row := range rows {
        table.Insert(ctx, row)

    return db

Then, you can connect to the server with any MySQL client:

> mysql --host= --port=3306 -u user -ppass test -e "SELECT * FROM mytable"
| name     | email             | phone_numbers                 | created_at          |
| John Doe |      | ["555-555-555"]               | 2018-04-18 10:42:58 |
| John Doe |   | []                            | 2018-04-18 10:42:58 |
| Jane Doe |      | []                            | 2018-04-18 10:42:58 |
| Evil Bob | | ["555-666-555","666-666-666"] | 2018-04-18 10:42:58 |

See the complete example here.

Queries examples

SELECT count(name) FROM mytable
| COUNT( |
|                   4 |

SELECT name,year(created_at) FROM mytable
| name     | YEAR(mytable.created_at) |
| John Doe |                     2018 |
| John Doe |                     2018 |
| Jane Doe |                     2018 |
| Evil Bob |                     2018 |

SELECT email FROM mytable WHERE name = 'Evil Bob'
| email             |
| |

Custom data source implementation

To create your own data source implementation you need to implement the following interfaces:

  • sql.Database interface. This interface will provide tables from your data source. You can also implement other interfaces on your database to unlock additional functionality:

    • sql.TableCreator to support creating new tables
    • sql.TableDropper to support dropping tables
    • sql.TableRenamer to support renaming tables
    • sql.ViewCreator to support creating persisted views on your tables
    • sql.ViewDropper to support dropping persisted views
  • sql.Table interface. This interface will provide rows of values from your data source. You can also implement other interfaces on your table to unlock additional functionality:

    • sql.InsertableTable to allow your data source to be updated with INSERT statements.
    • sql.UpdateableTable to allow your data source to be updated with UPDATE statements.
    • sql.DeletableTable to allow your data source to be updated with DELETE statements.
    • sql.ReplaceableTable to allow your data source to be updated with REPLACE statements.
    • sql.AlterableTable to allow your data source to have its schema modified by adding, dropping, and altering columns.
    • sql.IndexedTable to declare your table's native indexes to speed up query execution.
    • sql.IndexAlterableTable to accept the creation of new native indexes.
    • sql.ForeignKeyAlterableTable to signal your support of foreign key constraints in your table's schema and data.
    • sql.ProjectedTable to return rows that only contain a subset of the columns in the table. This can make query execution faster.
    • sql.FilteredTable to filter the rows returned by your table to those matching a given expression. This can make query execution faster (if your table implementation can filter rows more efficiently than checking an expression on every row in a table).

You can see a really simple data source implementation in the memory package.

Testing your data source implementation

go-mysql-server provides a suite of engine tests that you can use to validate that your implementation works as expected. See the enginetest package for details and examples.


go-mysql-server exposes a series of interfaces to allow you to implement your own indexes so you can speed up your queries.

Native indexes

Tables can declare that they support native indexes, which means that they support efficiently returning a subset of their rows that match an expression. The memory package contains an example of this behavior, but please note that it is only for example purposes and doesn't actually make queries faster (although we could change this in the future).

Integrators should implement the sql.IndexedTable interface to declare which indexes their tables support and provide a means of returning a subset of the rows based on an sql.IndexLookup provided by their sql.Index implementation. There are a variety of extensions to sql.Index that can be implemented, each of which unlocks additional capabilities:

  • sql.Index. Base-level interface, supporting equality lookups for an index.
  • sql.AscendIndex. Adds support for > and >= indexed lookups.
  • sql.DescendIndex. Adds support for < and <= indexed lookups.
  • sql.NegateIndex. Adds support for negating other index lookups.
  • sql.MergeableIndexLookup. Adds support for merging two sql.IndexLookups together to create a new one, representing AND and OR expressions on indexed columns.

Custom index driver implementation

Index drivers provide different backends for storing and querying indexes, without the need for a table to store and query its own native indexes. To implement a custom index driver you need to implement a few things:

  • sql.IndexDriver interface, which will be the driver itself. Not that your driver must return an unique ID in the ID method. This ID is unique for your driver and should not clash with any other registered driver. It's the driver's responsibility to be fault tolerant and be able to automatically detect and recover from corruption in indexes.
  • sql.Index interface, returned by your driver when an index is loaded or created.
  • sql.IndexValueIter interface, which will be returned by your sql.IndexLookup and should return the values of the index.
  • Don't forget to register the index driver in your sql.Context using context.RegisterIndexDriver(mydriver) to be able to use it.

To create indexes using your custom index driver you need to use extension syntax USING driverid on the index creation statement. For example:

CREATE INDEX foo ON table USING driverid (col1, col2)

go-mysql-server does not provide a production index driver implementation. We previously provided a pilosa implementation, but removed it due to the difficulty of supporting it on all platforms (pilosa doesn't work on Windows).

You can see an example of a driver implementation in the memory package.


go-mysql-server utilizes module to expose metrics (counters, gauges, histograms) for certain packages (so far for engine, analyzer, regex). If you already have metrics server (prometheus, statsd/statsite, influxdb, etc.) and you want to gather metrics also from go-mysql-server components, you will need to initialize some global variables by particular implementations to satisfy following interfaces:

// Counter describes a metric that accumulates values monotonically.
type Counter interface {
	With(labelValues ...string) Counter
	Add(delta float64)

// Gauge describes a metric that takes specific values over time.
type Gauge interface {
	With(labelValues ...string) Gauge
	Set(value float64)
	Add(delta float64)

// Histogram describes a metric that takes repeated observations of the same
// kind of thing, and produces a statistical summary of those observations,
// typically expressed as quantiles or buckets.
type Histogram interface {
	With(labelValues ...string) Histogram
	Observe(value float64)

You can use one of go-kit implementations or try your own. For instance, we want to expose metrics for prometheus server. So, before we start mysql engine, we have to set up the following variables:

    promopts ""


// engine metrics
sqle.QueryCounter = prometheus.NewCounterFrom(promopts.CounterOpts{
		Namespace: "go_mysql_server",
		Subsystem: "engine",
		Name:      "query_counter",
	}, []string{
sqle.QueryErrorCounter = prometheus.NewCounterFrom(promopts.CounterOpts{
    Namespace: "go_mysql_server",
    Subsystem: "engine",
    Name:      "query_error_counter",
}, []string{
sqle.QueryHistogram = prometheus.NewHistogramFrom(promopts.HistogramOpts{
    Namespace: "go_mysql_server",
    Subsystem: "engine",
    Name:      "query_histogram",
}, []string{

// analyzer metrics
analyzer.ParallelQueryCounter = prometheus.NewCounterFrom(promopts.CounterOpts{
    Namespace: "go_mysql_server",
    Subsystem: "analyzer",
    Name:      "parallel_query_counter",
}, []string{

// regex metrics
regex.CompileHistogram = prometheus.NewHistogramFrom(promopts.HistogramOpts{
    Namespace: "go_mysql_server",
    Subsystem: "regex",
    Name:      "compile_histogram",
}, []string{
regex.MatchHistogram = prometheus.NewHistogramFrom(promopts.HistogramOpts{
    Namespace: "go_mysql_server",
    Subsystem: "regex",
    Name:      "match_histogram",
}, []string{

One important note - internally we set some labels for metrics, that's why have to pass those keys like "duration", "query", "driver", ... when we register metrics in prometheus. Other systems may have different requirements.

Powered by go-mysql-server


go-mysql-server was originally developed by the {source-d} organzation, and this repository was originally forked from src-d. We want to thank the entire {source-d} development team for their work on this project, especially Miguel Molina (@erizocosmico) and Juanjo Álvarez Martinez (@juanjux).


Apache License 2.0, see LICENSE