Julia: limited printing of large arrays
As I mentioned in a comment, the way to do this in v0.5 is to use an IOContext
.
A very simple way to limit the data is to pass the :limit => true
parameter to the IOContext
;
julia> show(IOContext(stdout, :limit => true), v)[0.147959 0.414018 … 0.595528 0.852563; 0.32679 0.824953 … 0.432143 0.036279; … ; 0.877398 0.661854 … 0.197207 0.15596; 0.0522946 0.508075 … 0.835359 0.705987]
But this still does not print out the way the REPL does; that's because show
with two arguments uses a single-line display. To use a multiline display, pass text/plain
as the second argument to show
(a MIME type):
julia> show(IOContext(stdout, :limit => true), "text/plain", v)100×100 Array{Float64,2}: 0.147959 0.414018 0.0282934 … 0.816132 0.595528 0.852563 0.32679 0.824953 0.0582351 0.822526 0.432143 0.036279 0.754989 0.724317 0.533966 0.987273 0.931932 0.973622 0.547866 0.282694 0.0295411 0.75929 0.886218 0.0813057 0.0626663 0.111795 0.625083 0.439983 0.562143 0.669046 0.712093 0.469622 0.377298 … 0.298224 0.31853 0.376066 0.774625 0.754328 0.756725 0.61113 0.76566 0.999292 0.917846 0.308363 0.489246 0.715311 0.175302 0.124059 0.310922 0.140575 0.20635 0.0280192 0.683004 0.168129 0.753361 0.755103 0.831806 0.118009 0.122374 0.281476 ⋮ ⋱ 0.420264 0.7614 0.748408 0.330983 0.0776789 0.309464 0.984379 0.851735 0.595121 0.534982 0.255317 0.743713 0.814505 0.765941 0.71852 0.730677 0.477631 0.0360992 0.910384 0.0747604 0.490685 0.0904559 0.0756424 0.313898 0.628416 0.0790874 0.401488 … 0.523521 0.397249 0.58112 0.578361 0.336352 0.261118 0.838256 0.387374 0.451647 0.66724 0.586342 0.378968 0.602694 0.450686 0.901279 0.877398 0.661854 0.685156 0.658952 0.197207 0.15596 0.0522946 0.508075 0.244423 0.95935 0.835359 0.705987
You can of course change how many rows are shown by passing in :displaysize
to the IOContext
:
julia> show(IOContext(stdout, :limit => true, :displaysize => (10, 10)), "text/plain", v)100×100 Array{Float64,2}: 0.147959 … 0.852563 0.32679 0.036279 0.754989 0.973622 ⋮ ⋱ 0.877398 0.15596 0.0522946 0.705987
Overall, IOContext
is very flexible. See its documentation for more details.
You could easily define a head
function yourself. For one and two dimensions, this is pretty straightforward:
torange(n::Integer, m) = 1:min(n, m)torange(c::Colon, m) = (:)function head(a::AbstractArray{TypeVar(:T), 1}, n = 10) view(a, torange(n,size(a,1)))endfunction head(a::AbstractArray{TypeVar(:T), 2}, n1 = 10, n2 = 10) view(a, torange(n1, size(a,1)), torange(n2, size(a,2)))end
The torange
method allows to use a :
to return the full length in the corresponding dimension. For example
head(zeros(10, 10), 5, :)5×10 SubArray{Float64,2,Array{Float64,2},Tuple{UnitRange{Int64},Colon},false}: 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
For more than two dimensions, I have chosen to repeat the last argument beginning from the third:
function head{Na, Nn}(a::AbstractArray{TypeVar(:T), Na}, n1 = 10, n2 = 10, ns::Vararg{TypeVar(:T), Nn} = 2) nend = last(ns) view(a, torange(n1, size(a,1)), torange(n2, size(a,2)), (torange(ns[i], size(a,i+2)) for i = 1:Nn)..., (torange(nend, size(a, i)) for i in Nn+3:Na)...)end
For example:
head(rand(10, 10, 5, 5), 3, 3, 2) # the last two is the default value and can be omitted3×3×2×2 SubArray{Float64,4,Array{Float64,4},Tuple{UnitRange{Int64},UnitRange{Int64},UnitRange{Int64},UnitRange{Int64}},false}:[:, :, 1, 1] = 0.384724 0.7328 0.585211 0.738284 0.95145 0.362914 0.43928 0.94307 0.758541[:, :, 2, 1] = 0.78603 0.588877 0.677201 0.559547 0.800559 0.488433 0.993593 0.691884 0.236595[:, :, 1, 2] = 0.25732 0.90491 0.323905 0.300924 0.703919 0.813316 0.040522 0.776142 0.624097[:, :, 2, 2] = 0.746677 0.153574 0.155539 0.991624 0.90167 0.0880094 0.0423263 0.0153597 0.0608328
Note that the n-dimensional version is not type-stable, but that shouldn't matter for printing.
We can get the same output as REPL by using display()
(is this what you are looking for...?) Also, the "head" and "tail" parts can be printed by using array sections, e.g.,
disp( x ) = ( display(x) ; println() ; println() )A = diagm( [ i for i=1:100 ] )disp( A )disp( A[ 1:5, : ] ) # headdisp( A[ end-4:end, : ] ) # tailB = [ i for i=1:100 ]disp( B )disp( B[ 1:5 ] ) # headdisp( B[ end-4:end ] ) # tail
$ julia test.jl100x100 Array{Int64,2}: 1 0 0 0 0 0 0 0 0 0 0 0 … 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 6 0 0 0 0 0 0 … 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 7 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 8 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 9 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 10 0 0 0 0 0 0 0 0 0 0 0 ⋮ ⋮ ⋮ ⋱ ⋮ 0 0 0 0 0 0 0 0 0 0 0 0 92 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 93 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 94 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 95 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 … 0 0 0 0 96 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 97 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 98 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 99 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1005x100 Array{Int64,2}: 1 0 0 0 0 0 0 0 0 0 0 0 0 … 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 05x100 Array{Int64,2}: 0 0 0 0 0 0 0 0 0 0 0 0 0 … 0 0 0 0 0 96 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 97 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 98 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 99 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 100100-element Array{Int64,1}: 1 2 3 4 5 6 7 8 9 10 ⋮ 92 93 94 95 96 97 98 99 1005-element Array{Int64,1}: 1 2 3 4 55-element Array{Int64,1}: 96 97 98 99 100