Apache Arrowで行のシャッフル / Shuffles Rows in Apache Arrow Data

code:shuffled-table.rb

class ShuffledTable

def initialize(table)

@table = table

@indices = (0...n_rows).to_a.shuffle

end

def each_record

return to_enum(__method__) unless block_given?

@indices.each do |i|

yield Arrow::Record.new(@table, i)

end

end

end

if $PROGRAM_NAME == __FILE__

require "parquet"

table = Arrow::Table.load("path/to/data.parquet")

shuffled = ShuffledTable.new(table)

pp shuffled.each_record.take(3)

end