The (non-exported) Cartesian module provides macros that facilitate writing multidimensional algorithms. It is hoped that Cartesian will not, in the long term, be necessary; however, at present it is one of the few ways to write compact and performant multidimensional code.

Principles of usage

A simple example of usage is:


which generates the following code:


In general, Cartesian allows you to write generic code that contains repetitive elements, like the nested loops in this example. Other applications include repeated expressions (e.g., loop unwinding) or creating function calls with variable numbers of arguments without using the “splat” construct (i...).

Basic syntax

The (basic) syntax of @nloops is as follows:

  • The first argument must be an integer (not a variable) specifying the number of loops.
  • The second argument is the symbol-prefix used for the iterator variable. Here we used i, and variables i_1,i_2,i_3 were generated.
  • The third argument specifies the range for each iterator variable. If you use a variable (symbol) here, it’s taken as 1:size(A,dim). More flexibly, you can use the anonymous-function expression syntax described below.
  • The last argument is the body of the loop. Here, that’s what appears between the begin...end.

There are some additional features of @nloops described in the reference section.

@nref follows a similar pattern, generating A[i_1,i_2,i_3] from @nref3Ai. The general practice is to read from left to right, which is why @nloops is @nloops3iAexpr (as in fori_2=1:size(A,2), where i_2 is to the left and the range is to the right) whereas @nref is @nref3Ai (as in A[i_1,i_2,i_3], where the array comes first).

If you’re developing code with Cartesian, you may find that debugging is easier when you examine the generated code, using macroexpand:


Supplying the number of expressions

The first argument to both of these macros is the number of expressions, which must be an integer. When you’re writing a function that you intend to work in multiple dimensions, this may not be something you want to hard-code. If you’re writing code that you need to work with older Julia versions, currently you should use the @ngenerate macro described in an older version of this documentation.

Starting in Julia 0.4-pre, the recommended approach is to use a @generatedfunction. Here’s an example:

@generatedfunction mysum{T,N}(A::Array{T,N})quotes=zero(T)@nloops$NiAbegins+=@nref$NAiendsendend

Naturally, you can also prepare expressions or perform calculations before the quote block.

Anonymous-function expressions as macro arguments

Perhaps the single most powerful feature in Cartesian is the ability to supply anonymous-function expressions that get evaluated at parsing time. Let’s consider a simple example:


@nexprs generates n expressions that follow a pattern. This code would generate the following statements:


In each generated statement, an “isolated” j (the variable of the anonymous function) gets replaced by values in the range 1:2. Generally speaking, Cartesian employs a LaTeX-like syntax. This allows you to do math on the index j. Here’s an example computing the strides of an array:


would generate expressions


Anonymous-function expressions have many uses in practice.

Macro reference

@nloops N itersym rangeexpr bodyexpr
@nloops N itersym rangeexpr preexpr bodyexpr
@nloops N itersym rangeexpr preexpr postexpr bodyexpr

Generate N nested loops, using itersym as the prefix for the iteration variables. rangeexpr may be an anonymous-function expression, or a simple symbol var in which case the range is 1:size(var,d) for dimension d.

Optionally, you can provide “pre” and “post” expressions. These get executed first and last, respectively, in the body of each loop. For example, :


would generate :


If you want just a post-expression, supply nothing for the pre-expression. Using parenthesis and semicolons, you can supply multi-statement expressions.

@nref N A indexexpr

Generate expressions like A[i_1,i_2,...]. indexexpr can either be an iteration-symbol prefix, or an anonymous-function expression.

@nexprs N expr

Generate N expressions. expr should be an anonymous-function expression.

@ntuple N expr

Generates an N-tuple. @ntuple2i would generate (i_1,i_2), and @ntuple2k->k+1 would generate (2,3).

@nall N expr

@nall3d->(i_d>1) would generate the expression (i_1>1&&i_2>1&&i_3>1). This can be convenient for bounds-checking.

@nif N conditionexpr expr
@nif N conditionexpr expr elseexpr

Generates a sequence of if...elseif...else...end statements. For example:

@nif3d->(i_d>=size(A,d))d->(error("Dimension ",d," too big"))d->println("All OK")

would generate:

ifi_1>size(A,1)error("Dimension ",1," too big")elseifi_2>size(A,2)error("Dimension ",2," too big")elseprintln("All OK")end