Branch: master
Find file Copy path
Fetching contributors…
Cannot retrieve contributors at this time
155 lines (113 sloc) 4.55 KB
using JuliaDB


JuliaDB offers two main data structures as well as distributed counterparts. This allows you to easily scale up an analysis, as operations that work on non-distributed tables either work out of the box or are easy to transition for distributed tables.

Here is a high level overview of tables in JuliaDB:

  • Tables store data in columns.
  • Tables are typed.
    • Changing a table in some way therefore requires returning a new table (underlying data is not copied).
    • JuliaDB has few mutating operations because a new table is necessary in most cases.

Data for examples:

x = 1:10
y = vcat(fill('a', 4), fill('b', 6))
z = randn(10);


An IndexedTable is wrapper around a (named) tuple of Vectors, but it behaves like a Vector of (named) tuples. You can choose to sort the table by any number of primary keys (in this case columns :x and :y).

An IndexedTable is created with data in Julia via the table function or with data on disk via the loadtable function.

t = table((x=x, y=y, z=z); pkey = [:x, :y])


An NDSparse has a similar underlying structure to IndexedTable, but it behaves like a sparse array with arbitrary indices. The keys of an NDSparse are sorted, much like the primary keys of an IndexedTable.

An NDSparse is created with data in Julia via the ndsparse function or with data on disk via the loadndsparse function.

nd = ndsparse((x=x, y=y), (z=z,))
nd[1, 'a']
nd[10, 'j'].z
nd[1, :]


JuliaDB has a variety of ways to select columns. These selection methods get used across many JuliaDB's functions: select, reduce, groupreduce, groupby, join, pushcol, reindex, and more.

To demonstrate selection, we'll use the select function. A selection can be any of the following types:

  1. Integer -- returns the column at this position.
  2. Symbol -- returns the column with this name.
  3. Pair{Selection => Function} -- selects and maps a function over the selection, returns the result.
  4. AbstractArray -- returns the array itself. This must be the same length as the table.
  5. Tuple of Selection -- returns a table containing a column for every selector in the tuple.
  6. Regex -- returns the columns with names that match the regular expression.
  7. Type -- returns columns with elements of the given type.
  8. Not(Selection) -- returns columns that are not included in the selection.
  9. Between(first, last) -- returns columns between first and last.
  10. Keys() -- return the primary key columns.
t = table(1:10, randn(10), rand(Bool, 10); names = [:x, :y, :z])

select the :x vector

select(t, 1)
select(t, :x)

map a function to the :y vector

select(t, 2 => abs)
select(t, :y => x -> x > 0 ? x : -x)

select the table of :x and :z

select(t, (:x, :z))
select(t, r"(x|z)")

map a function to the table of :x and :y

select(t, (:x, :y) => row -> row[1] + row[2])
select(t, (1, :y) => row -> row.x + row.y)

select columns that are subtypes of Integer

select(t, Integer)

select columns that are not subtypes of Integer

select(t, Not(Integer))

Loading and Saving

using Pkg

Loading Data From CSV

Loading a CSV file (or multiple files) into one of JuliaDB's tabular data structures is accomplished via the loadtable and loadndsparse functions.

using JuliaDB, DelimitedFiles

x = rand(10, 2)
writedlm("temp.csv", x, ',')

t = loadtable("temp.csv")

!!! note loadtable and loadndsparse use Missing to represent missing values. To load a CSV that instead uses DataValue, see CSVFiles.jl. For more information on missing value representations, see Missing Values.

Converting From Other Data Structures

using JuliaDB, RDatasets

df = dataset("datasets", "iris")  # load data as DataFrame

table(df)  # Convert DataFrame to IndexedTable

Save Table into Binary Format

A table can be saved to disk (for fast, efficient reloading) via the save function.

Load Table from Binary Format

Tables that have been save-ed can be loaded efficiently via load.