CODEWITHSUNDEEP
arraysjulia
hr0m
hr0m

Reputation: 2805

How do I sum over a dimension in array while dropping the traversed dimension?

I want to do a sum over one dimension in an array. That's easy. For an array 9x100x100:

sum(a,1)

However what is then left is an array with dimension 1x100x100. And now I want to get rid of the first dimension, since there is only one element left. So my solution is just:

reshape(summed_array, 100,100)

In order to get the 100x100 array, which I wanted. However this does not feel very clean. Is there a better way of achieving this?

Upvotes: 7

Views: 3052

Answers (1)

jub0bs
jub0bs

Reputation: 66244

Update

As @E_net4 points out in a comment below, since Julia 1.0, you should use dropdims (a much better name!) rather than squeeze.

Original answer

You're looking for squeeze:

squeeze(A, dims)

Remove the dimensions specified by dims from array A. Elements of dims must be unique and within the range 1:ndims(A).

Example

julia> a = rand(4,3,2)
4x3x2 Array{Float64,3}:
[:, :, 1] =
 0.333543  0.83446   0.659689
 0.927134  0.885299  0.909313
 0.183557  0.263095  0.741925
 0.744499  0.509219  0.570718

[:, :, 2] =
 0.967247  0.90947   0.715283
 0.659315  0.667984  0.168867
 0.120959  0.842117  0.217277
 0.516499  0.60886   0.616639

julia> b = sum(a, 1)
1x3x2 Array{Float64,3}:
[:, :, 1] =
 2.18873  2.49207  2.88165

[:, :, 2] =
 2.26402  3.02843  1.71807

julia> c = squeeze(b, 1)
3x2 Array{Float64,2}:
 2.18873  2.26402
 2.49207  3.02843
 2.88165  1.71807

Upvotes: 11

Related Questions