Julia Reference Card (Page 2 of 2) in pdf

1.6 Arrays

function f(params) ... end

function definition

readsandwrite(command)

(its in, its out, process)

(homogeneous & type may be specified)/

f(x, y=0) = return x+y

optional parameter

detach(command)

run & outlive Julia process

Array(T, dims)

Uninitalized dense dim

f(a,b,c...) =

varargs c = [] or ()

setenv(command,env)

set vars for running

array of Type T

f(a,b; dir="nth") =

named args

ENV

EnvHash->EnvHash

Sys environment vars

Initialize different arrays (sometimes with T else just dims):

f(a,b; d=5,e...) =

varargs as of k,v tuples

getpid()

get Julias pid

zeros ones

trues

falses

rand randdf

randn

... |> (x,f)

apply fn to preceding args

clipboard(x)

print x to clipboard

eye

eye(n)

linspace(start, stop, n)

f(1,1), f(2), f(y=3, x=4)

function calls

s = clipboard()

or s from clipboard

Vector = 1 dim column array or cell is like a list in Python

(x) -> x+a

anonymous function

@time()

@elapsed()->secs

time and expression

Functions on arrays:

strftime([f,]time())

time as string

nnz (num non zero values) stride(A,n)

strides(A)

function make_adder_2(a)

closure, alternatively,

cd(f[,dir])

run f in temporary dir

ndims, transpose & ctranspose .' & '

add(b) = return a+b

function mkadr(a)

cell(dims ...)

uninitialized

return add

b -> a+b

Filesystem Operations

heterogeneous array

end

gethostname(), getipaddr(), pwd(), cd('dir'),

Array(Int32,0)

[]

{}

Empty array or cell

let v=1,w=" "... ; end

scope block with vars

mkdir(p,mode), mkpath(p,mode), rmdir(p),

reshape(A, dims)

new shape, same data

@time()

gc_disable()

report time elapsed

ignorestatus(cmd),

similar

reinterpret

@profile

Profile.print

profile, print & clear

redirect in run commands: |> std output, |>> append

[a,x] [a x]

{c,x} {c x}

add element

stdout, |.> stderr

to process, file or DevNull

a = [f(x,y,...) for x=rx,

array comprehensions

global v

bind to global variable

6 Input/Output

y=ry, ...]

: ... ; quote ... end

create an expression

1.7 DataFrames

(using DataFrames)

open(filename, mode)

open file (a & w create,

eval(expr)

evaluate expression

DataArray

NAtype

Array with missing values

mode = r r+ w w+ a a+

+ is both r&w, w truncates)

DataFrame(A=1:4,B=[...])

Tabular hetero dataset

open(f(),args)

f(result of open args)

using name

fn()

load module namespace

removeNA

replaceNA(dv,val)

remove or replace NAs

close(stream)

flush and close

import name

name.fn()

import gives named access

failNA

write(stream, x)

write binary x to stream

require(filepath)

Load file once

df[2,"A"]

df[[rows,]1:2]

get & slice

writedlm writecsv

array, dlm with csv delimeter

reload(filepath)

and reload it

df[1:2,["A","B"]]

read(stream, type[,dims]) read value from stream

include(filepath)

set dir & load source

df[df["A"] % 2 .== 0, :]

readbytes readdlm readcsv nb bytes, array, csv

evalfile(file)

execute file

colnames!(df[,newnames])

read or insert col names

readall readline[s]

all as string, line or lines

cd("data") do

safely write file in a

head, tail, describe

position(s)

get position of a stream

open("outfile", "w") do f

directory and close after.

join(df..., jointype)

join two dataframes

seek(s, pos)

seek stream to position

write(f,data)

groupby(df,catvar)

split df by categorical var

seekstart(s),

seekend(s)

to start to end

end

by(df,catvar, df->f(df[]))

split and apply fn or

skip(s, offset)

seek relative

end

by(df,catvar,:(n=...; m=...))

expression to subset

isopen

eof

isreadonly

open? end of file? read only?

4 Exception Handling & Debugging

stack(df,categorical var)

reshape data

ltoh(x)

htol(x)

little endian conversions

readtable(fname,header=false,

defaults are true, comma

try ...

Try-block

[de]serialize(stream,val)

separator='\t')

also, writetable(f, fname)

catch [y]

catch exception as var

download(url[,localfile]) unix download

2 Basic Syntax

print data

+ others

error("...")

exception handling

if expr statements

Conditional

end

7 Areas Not Covered

elseif expr statements

; if on same line

finally ...

in any case

else statements end

terminate with end

@assert expression

debug assertion

z = cond ? x : y

ternary version

Julia has a dynamic type system but with a rich language of

throw(e)

explicit expection

z = ifelse(cond, x, y)

as ? but all args evaluated

types including parametric. Type declaration is optional so

using Debug

loads the debugger

z = cond && x

short circuit

the casual user can usually ignore it.

@debug

@bp

before module, set breakpt

Multiple dispatch permits methods to be called based on the

p var1, ...

list lines, print vars

while expr statements end

while loop

types of all unnamed arguments.

step into, continue to @bp

while true .. if cond break do .. while equivalent

Object orientated design can be achieved by combining type

5 System Interaction

for target in iter

for loop

definition and multiple dispatch to associate methods with

statements ; end

run(`cmd`) or ;cmd

system call

new classes of objects

for i=itr, j=itr ... end

over multiple variables

spawn(`cmd`)

run asynchronously

Tasks or Coroutines permit computations to be flexibly

for key in keys(d)...

over dictionary

success(`cmd`)

bool for exit condition

suspended and resumed, effectively enhanced generators.

break, continue

end loop / jump to next

process_running(process)

determine if running

Parallel and distributed computing and metaprogramming

s=start(I);while !done(I,s)

iterator from sequence

process_exited(process)

determine if has exited

are supported.

(i,s) = next(I,s)

kill(process, signum)

print("hello world")

print or println (new line)

readsfrom(command)

(its stdout, process)

help(name)

get help on object

[expr for x in seq lc ]

list comprehension

writesto(command)

runs asynch & returns

apropos("search string")

search docs for string

nothing

empty statement

(its stdin, process)

Julia Reference Card Page 2

Related Articles

Related forms

Related Categories