Arbitrary row/column selection for NMatrix. Allows you to do selections like:
x.col([0,2]).row(y =~ 'setosa')
Works with dense and sparse matrix types.
gem install nmatrix-rowcol
require 'nmatrix/rowcol'
m = NMatrix.indgen [3,3]
# [ [0, 1, 2]
# [3, 4, 5]
# [6, 7, 8] ]
m.row [0,2]
# [ [0, 1, 2]
# [6, 7, 8] ]
m.row m.sum(1) > 6
# [ [3, 4, 5]
# [6, 7, 8] ]
m.col [0,2]
# [ [0, 2]
# [3, 5]
# [6, 8] ]
m.col m.sum(0) > 12
# [ [2]
# [5]
# [8] ]
m.row([0,2]).col([0,2])
# [ [0, 2]
# [6, 8] ]
m.col [2,0,1] # re-arrange columns
# [ [2, 0, 1]
# [5, 3, 4]
# [8, 6, 7] ]
m.row m.rows.times.to_a.sample(2) # random row samples
# [ [6, 7, 8]
# [0, 1, 2] ]MIT