379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
|
# File 'lib/rust/core/types/dataframe.rb', line 379
def aggregate(by, **aggregators)
raise TypeError, "Expected a string" unless by.is_a?(String)
raise TypeError, "All the aggregators should be procs" unless aggregators.values.all? { |v| v.is_a?(Proc) }
raise "Expected a block for default aggregator" unless block_given?
aggregators = aggregators.map { |label, callable| [label.to_s, callable] }.to_h
sorted = self.sort_by(by)
current_value = nil
partials = []
partial = nil
sorted.column(by).each_with_index do |value, index|
if current_value != value
current_value = value
partials << partial if partial
partial = Rust::DataFrame.new(self.column_names)
end
partial << sorted.fast_row(index)
end
partials << partial
result = Rust::DataFrame.new(self.column_names)
partials.each do |partial|
aggregated_row = {}
aggregated_row[by] = partial.column(by)[0]
(self.column_names - [by]).each do |column|
if aggregators[column]
aggregated_row[column] = aggregators[column].call(partial.column(column))
else
aggregated_row[column] = yield partial.column(column)
end
end
result << aggregated_row
end
return result
end
|