logo

Methods for Calculating Arrays Column-wise in Julia 📂Julia

Methods for Calculating Arrays Column-wise in Julia

Code

There are methods like map, broadcast, and comprehension for calculating arrays by column or row.

julia> using Statistics

julia> X = stack([i*ones(8) for i ∈ 1:9], dims=2)
8×9 Matrix{Float64}:
 1.0  2.0  3.0  4.0  5.0  6.0  7.0  8.0  9.0
 1.0  2.0  3.0  4.0  5.0  6.0  7.0  8.0  9.0
 1.0  2.0  3.0  4.0  5.0  6.0  7.0  8.0  9.0
 1.0  2.0  3.0  4.0  5.0  6.0  7.0  8.0  9.0
 1.0  2.0  3.0  4.0  5.0  6.0  7.0  8.0  9.0
 1.0  2.0  3.0  4.0  5.0  6.0  7.0  8.0  9.0
 1.0  2.0  3.0  4.0  5.0  6.0  7.0  8.0  9.0
 1.0  2.0  3.0  4.0  5.0  6.0  7.0  8.0  9.0

# map
julia> map(mean, eachcol(X))
9-element Vector{Float64}:
 1.0
 2.0
 3.0
 4.0
 5.0
 6.0
 7.0
 8.0
 9.0

julia> map(eachcol(X)) do col
           mean(col)
       end
9-element Vector{Float64}:
 1.0
 2.0
 3.0
 4.0
 5.0
 6.0
 7.0
 8.0
 9.0

# 브로드캐스트
julia> mean.(eachcol(X))
9-element Vector{Float64}:
 1.0
 2.0
 3.0
 4.0
 5.0
 6.0
 7.0
 8.0
 9.0

# 컴프리헨션
julia> [mean(col) for col ∈ eachcol(X)]
9-element Vector{Float64}:
 1.0
 2.0
 3.0
 4.0
 5.0
 6.0
 7.0
 8.0
 9.0

Differences1

Generally, there is no performance difference between map and comprehension. The difference lies in the returned value. Map returns a tuple when the iterator is given as a tuple.

julia> X = (1, 2, 3, 4, 5)
(1, 2, 3, 4, 5)

julia> [x^2 for x ∈ X]
5-element Vector{Int64}:
  1
  4
  9
 16
 25

julia> map(x->x^2, X)
(1, 4, 9, 16, 25)

To obtain a tuple using comprehension,

julia> (x^2 for x ∈ X).iter
(1, 2, 3, 4, 5)

Environment

  • OS: Windows11
  • Version: Julia 1.10.0

  1. Bogumił Kamiński, 데이터 분석을 위한 줄리아(Julia for Data Analysis, 류현지 역), p92 ↩︎