Class: Arrow::Table

Inherits:
Object
  • Object
show all
Includes:
ColumnContainable, GenericFilterable, GenericTakeable, InputReferable, RecordContainable
Defined in:
lib/arrow/table.rb

Class Method Summary collapse

Instance Method Summary collapse

Methods included from RecordContainable

#each_record

Methods included from InputReferable

#refer_input, #share_input

Methods included from GenericTakeable

included, #take_generic

Methods included from GenericFilterable

#filter_generic, included

Methods included from ColumnContainable

#[], #column_names, #columns, #each_column, #find_column, #select_columns

Constructor Details

#initialize(columns) ⇒ Table #initialize(raw_table) ⇒ Table #initialize(raw_table) ⇒ Table #initialize(raw_table) ⇒ Table #initialize(schema, columns) ⇒ Table #initialize(schema, arrays) ⇒ Table #initialize(schema, record_batches) ⇒ Table #initialize(schema, raw_records) ⇒ Table

Creates a new Arrow::Table.

Overloads:

  • #initialize(columns) ⇒ Table

    Examples:

    Create a table from columns

    count_field = Arrow::Field.new("count", :uint32)
count_array = Arrow::UInt32Array.new([0, 2, nil, 4])
count_column = Arrow::Column.new(count_field, count_array)
visible_field = Arrow::Field.new("visible", :boolean)
visible_array = Arrow::BooleanArray.new([true, nil, nil, false])
visible_column = Arrow::Column.new(visible_field, visible_array)
Arrow::Table.new([count_column, visible_column])

    Parameters:

  • #initialize(raw_table) ⇒ Table

    Examples:

    Create a table from column name and values

    Arrow::Table.new("count" => Arrow::UInt32Array.new([0, 2, nil, 4]),
                 "visible" => Arrow::BooleanArray.new([true, nil, nil, false]))

    Parameters:

    • raw_table (Hash<String, Arrow::Array>) —

      The pairs of column name and values of the table. Column values is ‘Arrow::Array`.

  • #initialize(raw_table) ⇒ Table

    Examples:

    Create a table from column name and values

    count_chunks = [
  Arrow::UInt32Array.new([0, 2]),
  Arrow::UInt32Array.new([nil, 4]),
]
visible_chunks = [
  Arrow::BooleanArray.new([true]),
  Arrow::BooleanArray.new([nil, nil, false]),
]
Arrow::Table.new("count" => Arrow::ChunkedArray.new(count_chunks),
                 "visible" => Arrow::ChunkedArray.new(visible_chunks))

    Parameters:

    • raw_table (Hash<String, Arrow::ChunkedArray>) —

      The pairs of column name and values of the table. Column values is ‘Arrow::ChunkedArray`.

  • #initialize(raw_table) ⇒ Table

    Examples:

    Create a table from column name and values

    Arrow::Table.new("count" => [0, 2, nil, 4],
                 "visible" => [true, nil, nil, false])

    Parameters:

    • raw_table (Hash<String, ::Array>) —

      The pairs of column name and values of the table. Column values is ‘Array`.

  • #initialize(schema, columns) ⇒ Table

    Examples:

    Create a table from schema and columns

    count_field = Arrow::Field.new("count", :uint32)
count_array = Arrow::UInt32Array.new([0, 2, nil, 4])
count_column = Arrow::Column.new(count_field, count_array)
visible_field = Arrow::Field.new("visible", :boolean)
visible_array = Arrow::BooleanArray.new([true, nil, nil, false])
visible_column = Arrow::Column.new(visible_field, visible_array)
Arrow::Table.new(Arrow::Schema.new([count_field, visible_field]),
                 [count_column, visible_column])

    Parameters:

  • #initialize(schema, arrays) ⇒ Table

    Examples:

    Create a table from schema and arrays

    count_field = Arrow::Field.new("count", :uint32)
count_array = Arrow::UInt32Array.new([0, 2, nil, 4])
visible_field = Arrow::Field.new("visible", :boolean)
visible_array = Arrow::BooleanArray.new([true, nil, nil, false])
Arrow::Table.new(Arrow::Schema.new([count_field, visible_field]),
                 [count_array, visible_array])

    Parameters:

  • #initialize(schema, record_batches) ⇒ Table

    Examples:

    Create a table from schema and record batches

    count_field = Arrow::Field.new("count", :uint32)
visible_field = Arrow::Field.new("visible", :boolean)
schema = Arrow::Schema.new([count_field, visible_field])
record_batches = [
  Arrow::RecordBatch.new(schema, [[0, true], [2, nil], [nil, nil]]),
  Arrow::RecordBatch.new(schema, [[4, false]]),
]
Arrow::Table.new(schema, record_batches)

    Parameters:

  • #initialize(schema, raw_records) ⇒ Table

    Examples:

    Create a table from schema and raw records

    schema = {
  count: :uint32,
  visible: :boolean,
}
raw_records = [
  [0, true],
  [2, nil],
  [nil, nil],
  [4, false],
]
Arrow::Table.new(schema, raw_records)

    Parameters:

    • schema (Arrow::Schema) —

      The schema of the table. You can also specify schema as primitive Ruby objects. See Schema#initialize for details.

    • raw_records (::Array<::Array>) —

      The data of the table as primitive Ruby objects.



163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
 
# File 'lib/arrow/table.rb', line 163

def initialize(*args)
  n_args = args.size
  case n_args
  when 1
    raw_table_converter = RawTableConverter.new(args[0])
    schema = raw_table_converter.schema
    values = raw_table_converter.values
  when 2
    schema = args[0]
    schema = Schema.new(schema) unless schema.is_a?(Schema)
    values = args[1]
    case values[0]
    when ::Array
      values = [RecordBatch.new(schema, values)]
    when Column
      values = values.collect(&:data)
    end
  else
    message = "wrong number of arguments (given #{n_args}, expected 1..2)"
    raise ArgumentError, message
  end
  initialize_raw(schema, values)
end

Dynamic Method Handling

This class handles dynamic methods through the method_missing method

#method_missing(name, *args, &block) ⇒ Object 



622
623
624
625
626
627
628
 
# File 'lib/arrow/table.rb', line 622

def method_missing(name, *args, &block)
  if args.empty?
    column = find_column(name)
    return column if column
  end
  super
end

Class Method Details

.load(path, options = {}) ⇒ Object 



29
30
31
 
# File 'lib/arrow/table.rb', line 29

def load(path, options={})
  TableLoader.load(path, options)
end

Instance Method Details

#each_record_batch ⇒ Object 



187
188
189
190
191
192
193
194
195
 
# File 'lib/arrow/table.rb', line 187

def each_record_batch
  return to_enum(__method__) unless block_given?

  reader = TableBatchReader.new(self)
  while record_batch = reader.read_next
    share_input(record_batch)
    yield(record_batch)
  end
end

#group(*keys) ⇒ Object

Experimental



436
437
438
 
# File 'lib/arrow/table.rb', line 436

def group(*keys)
  Group.new(self, keys)
end

#inspect ⇒ Object 



613
614
615
 
# File 'lib/arrow/table.rb', line 613

def inspect
  "#{super}\n#{to_s}"
end

#inspect_raw ⇒ Object 



612
 
# File 'lib/arrow/table.rb', line 612

alias_method :inspect_raw, :inspect

#join(right, type: :inner, left_outputs: nil, right_outputs: nil) ⇒ Arrow::Table #join(right, key, type: :inner, left_outputs: nil, right_outputs: nil) ⇒ Arrow::Table #join(right, keys, type: :inner, left_suffix: "", right_suffix: "") ⇒ Arrow::Table #join(right, keys, type: :inner, left_outputs: nil, right_outputs: nil) ⇒ Arrow::Table

Join another Table by matching with keys.

Join columns with ‘right` on join key columns.

Join right by keys.

Column name can be renamed by appending `left_suffix` or `right_suffix`.

@macro join_common_before
@param keys [::Array<String, Symbol>] Join keys.
@macro join_common_after

Overloads:

  • #join(right, type: :inner, left_outputs: nil, right_outputs: nil) ⇒ Arrow::Table

    If key(s) are not supplied, common keys in self and right are used (natural join).

    Column used as keys are merged and remain in left side when both of ‘left_outputs` and `right_outputs` are `nil`.

    Join columns with ‘right` on join key columns.

    Parameters:

    • right (Arrow::Table) —

      The right table.

    • type (Arrow::JoinType) (defaults to: :inner) —

      How to join.

    • left_outputs (::Array<String, Symbol>) (defaults to: nil) —

      Output columns in ‘self`.

      If both of ‘left_outputs` and `right_outputs` aren’t specified, all columns in ‘self` and `right` are output.

    • right_outputs (::Array<String, Symbol>) (defaults to: nil) —

      Output columns in ‘right`.

      If both of ‘left_outputs` and `right_outputs` aren’t specified, all columns in ‘self` and `right` are output.

    Returns:

  • #join(right, key, type: :inner, left_outputs: nil, right_outputs: nil) ⇒ Arrow::Table

    Join right by a key.

    Column used as keys are merged and remain in left side when both of ‘left_outputs` and `right_outputs` are `nil`.

    Join columns with ‘right` on join key columns.

    Parameters:

    • right (Arrow::Table) —

      The right table.

    • key (String, Symbol) —

      A join key.

    • type (Arrow::JoinType) (defaults to: :inner) —

      How to join.

    • left_outputs (::Array<String, Symbol>) (defaults to: nil) —

      Output columns in ‘self`.

      If both of ‘left_outputs` and `right_outputs` aren’t specified, all columns in ‘self` and `right` are output.

    • right_outputs (::Array<String, Symbol>) (defaults to: nil) —

      Output columns in ‘right`.

      If both of ‘left_outputs` and `right_outputs` aren’t specified, all columns in ‘self` and `right` are output.

    Returns:

  • #join(right, keys, type: :inner, left_suffix: "", right_suffix: "") ⇒ Arrow::Table

    left_outputs: nil, right_outputs: nil)

  • #join(right, keys, type: :inner, left_outputs: nil, right_outputs: nil) ⇒ Arrow::Table

    Join right by a key or keys mapped by a hash.

    Join columns with ‘right` on join key columns.

    Parameters:

    • right (Arrow::Table) —

      The right table.

    • keys (Hash) —

      Specify join keys in ‘self` and `right` separately.

    • type (Arrow::JoinType) (defaults to: :inner) —

      How to join.

    • left_outputs (::Array<String, Symbol>) (defaults to: nil) —

      Output columns in ‘self`.

      If both of ‘left_outputs` and `right_outputs` aren’t specified, all columns in ‘self` and `right` are output.

    • right_outputs (::Array<String, Symbol>) (defaults to: nil) —

      Output columns in ‘right`.

      If both of ‘left_outputs` and `right_outputs` aren’t specified, all columns in ‘self` and `right` are output.

    Options Hash (keys):

    • :left (String, Symbol, ::Array<String, Symbol>) —

      Join keys in ‘self`.

    • :right (String, Symbol, ::Array<String, Symbol>) —

      Join keys in ‘right`.

    Returns:

Parameters:

  • right (Arrow::Table) —

    The right table.

  • type (Arrow::JoinType) (defaults to: :inner) —

    How to join.

  • left_outputs (::Array<String, Symbol>) (defaults to: nil) —

    Output columns in ‘self`.

    If both of ‘left_outputs` and `right_outputs` aren’t specified, all columns in ‘self` and `right` are output.

  • right_outputs (::Array<String, Symbol>) (defaults to: nil) —

    Output columns in ‘right`.

    If both of ‘left_outputs` and `right_outputs` aren’t specified, all columns in ‘self` and `right` are output.

Returns:

Since:

  • 11.0.0

  • 7.0.0



527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
 
# File 'lib/arrow/table.rb', line 527

def join(right,
         keys=nil,
         type: :inner,
         left_suffix: "",
         right_suffix: "",
         left_outputs: nil,
         right_outputs: nil)
  is_natural_join = keys.nil?
  keys ||= (column_names & right.column_names)
  type = JoinType.try_convert(type) || type
  plan = ExecutePlan.new
  left_node = plan.build_source_node(self)
  right_node = plan.build_source_node(right)
  if keys.is_a?(Hash)
    left_keys = keys[:left]
    right_keys = keys[:right]
  else
    left_keys = keys
    right_keys = keys
  end
  left_keys = Array(left_keys)
  right_keys = Array(right_keys)
  hash_join_node_options = HashJoinNodeOptions.new(type,
                                                   left_keys,
                                                   right_keys)
  use_manual_outputs = false
  unless left_outputs.nil?
    hash_join_node_options.left_outputs = left_outputs
    use_manual_outputs = true
  end
  unless right_outputs.nil?
    hash_join_node_options.right_outputs = right_outputs
    use_manual_outputs = true
  end
  hash_join_node = plan.build_hash_join_node(left_node,
                                             right_node,
                                             hash_join_node_options)
  type_nick = type.nick
  is_filter_join = (type_nick.end_with?("-semi") or
                    type_nick.end_with?("-anti"))
  if use_manual_outputs or is_filter_join
    process_node = hash_join_node
  elsif is_natural_join
    process_node = join_merge_keys(plan, hash_join_node, right, keys)
  elsif keys.is_a?(String) or keys.is_a?(Symbol)
    process_node = join_merge_keys(plan, hash_join_node, right, [keys.to_s])
  elsif !keys.is_a?(Hash) and (left_suffix != "" or right_suffix != "")
    process_node = join_rename_keys(plan,
                                    hash_join_node,
                                    right,
                                    keys,
                                    left_suffix,
                                    right_suffix)
  else
    process_node = hash_join_node
  end
  sink_node_options = SinkNodeOptions.new
  plan.build_sink_node(process_node, sink_node_options)
  plan.validate
  plan.start
  plan.wait
  reader = sink_node_options.get_reader(process_node.output_schema)
  table = reader.read_all
  share_input(table)
  table
end

#merge(other) ⇒ Arrow::Table

TODO

Returns:



360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
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
 
# File 'lib/arrow/table.rb', line 360

def merge(other)
  added_columns = {}
  removed_columns = {}

  case other
  when Hash
    other.each do |name, value|
      name = name.to_s
      if value
        added_columns[name] = ensure_raw_column(name, value)
      else
        removed_columns[name] = true
      end
    end
  when Table
    added_columns = {}
    other.columns.each do |column|
      name = column.name
      added_columns[name] = ensure_raw_column(name, column)
    end
  else
    message = "merge target must be Hash or Arrow::Table: " +
      "<#{other.inspect}>: #{inspect}"
    raise ArgumentError, message
  end

  new_columns = []
  columns.each do |column|
    column_name = column.name
    new_column = added_columns.delete(column_name)
    if new_column
      new_columns << new_column
      next
    end
    next if removed_columns.key?(column_name)
    new_columns << ensure_raw_column(column_name, column)
  end
  added_columns.each do |name, new_column|
    new_columns << new_column
  end
  new_fields = []
  new_arrays = []
  new_columns.each do |new_column|
    new_fields << new_column[:field]
    new_arrays << new_column[:data]
  end
  table = self.class.new(new_fields, new_arrays)
  share_input(table)
  table
end

#pack ⇒ Object 



450
451
452
453
454
455
456
457
 
# File 'lib/arrow/table.rb', line 450

def pack
  packed_arrays = columns.collect do |column|
    column.data.pack
  end
  table = self.class.new(schema, packed_arrays)
  share_input(table)
  table
end

#remove_column(name_or_index) ⇒ Object 



412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
 
# File 'lib/arrow/table.rb', line 412

def remove_column(name_or_index)
  case name_or_index
  when String, Symbol
    name = name_or_index.to_s
    index = columns.index {|column| column.name == name}
    if index.nil?
      message = "unknown column: #{name_or_index.inspect}: #{inspect}"
      raise KeyError.new(message)
    end
  else
    index = name_or_index
    index += n_columns if index < 0
    if index < 0 or index >= n_columns
      message = "out of index (0..#{n_columns - 1}): " +
        "#{name_or_index.inspect}: #{inspect}"
      raise IndexError.new(message)
    end
  end
  table = remove_column_raw(index)
  share_input(table)
  table
end

#remove_column_raw ⇒ Object 



411
 
# File 'lib/arrow/table.rb', line 411

alias_method :remove_column_raw, :remove_column

#respond_to_missing?(name, include_private) ⇒ Boolean

Returns:

  • (Boolean) 


617
618
619
620
 
# File 'lib/arrow/table.rb', line 617

def respond_to_missing?(name, include_private)
  return true if find_column(name)
  super
end

#save(output, options = {}) ⇒ Object 



445
446
447
448
 
# File 'lib/arrow/table.rb', line 445

def save(output, options={})
  saver = TableSaver.new(self, output, options)
  saver.save
end

#slice(offset, length) ⇒ Arrow::Table #slice(index) ⇒ Arrow::Record #slice(booleans) ⇒ Arrow::Table #slice(boolean_array) ⇒ Arrow::Table #slice(range) ⇒ Arrow::Table #slice(conditions) ⇒ Arrow::Table #slice {|slicer| ... } ⇒ Arrow::Table

Overloads:

  • #slice(offset, length) ⇒ Arrow::Table

    Returns The sub ‘Arrow::Table` that covers only from `offset` to `offset + length` range.

    Parameters:

    • offset (Integer) —

      The offset of sub Arrow::Table.

    • length (Integer) —

      The length of sub Arrow::Table.

    Returns:

    • (Arrow::Table) —

      The sub ‘Arrow::Table` that covers only from `offset` to `offset + length` range.

  • #slice(index) ⇒ Arrow::Record

    Returns The ‘Arrow::Record` corresponding to index of the table.

    Parameters:

    • index (Integer) —

      The index in this table.

    Returns:

    • (Arrow::Record) —

      The ‘Arrow::Record` corresponding to index of the table.

  • #slice(booleans) ⇒ Arrow::Table

    Returns The sub ‘Arrow::Table` that covers only rows of indices the values of `booleans` is true.

    Parameters:

    • booleans (::Array<Boolean>) —

      The values indicating the target rows.

    Returns:

    • (Arrow::Table) —

      The sub ‘Arrow::Table` that covers only rows of indices the values of `booleans` is true.

  • #slice(boolean_array) ⇒ Arrow::Table

    Returns The sub ‘Arrow::Table` that covers only rows of indices the values of `boolean_array` is true.

    Parameters:

    • boolean_array (::Array<Arrow::BooleanArray>) —

      The values indicating the target rows.

    Returns:

    • (Arrow::Table) —

      The sub ‘Arrow::Table` that covers only rows of indices the values of `boolean_array` is true.

  • #slice(range) ⇒ Arrow::Table

    Returns The sub ‘Arrow::Table` that covers only rows of the range of indices.

    Parameters:

    • range_included_end (Range) —

      The range indicating the target rows.

    Returns:

    • (Arrow::Table) —

      The sub ‘Arrow::Table` that covers only rows of the range of indices.

  • #slice(conditions) ⇒ Arrow::Table

    Returns The sub ‘Arrow::Table` that covers only rows matched by condition.

    Parameters:

    • conditions (Hash) —

      The conditions to select records.

    Returns:

    • (Arrow::Table) —

      The sub ‘Arrow::Table` that covers only rows matched by condition

  • #slice {|slicer| ... } ⇒ Arrow::Table

    Returns The sub ‘Arrow::Table` that covers only rows matched by condition specified by slicer.

    Yields:

    • (slicer) —

      Gives slicer that constructs condition to select records.

    Yield Parameters:

    • slicer (Arrow::Slicer) —

      The slicer that helps us to build condition.

    Yield Returns:

    Returns:

    • (Arrow::Table) —

      The sub ‘Arrow::Table` that covers only rows matched by condition specified by slicer.



255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
 
# File 'lib/arrow/table.rb', line 255

def slice(*args)
  slicers = []
  if block_given?
    unless args.empty?
      raise ArgumentError, "must not specify both arguments and block"
    end
    block_slicer = yield(Slicer.new(self))
    case block_slicer
    when ::Array
      slicers.concat(block_slicer)
    else
      slicers << block_slicer
    end
  else
    expected_n_args = nil
    case args.size
    when 1
      case args[0]
      when Integer
        index = args[0]
        index += n_rows if index < 0
        return nil if index < 0
        return nil if index >= n_rows
        return Record.new(self, index)
      when Hash
        condition_pairs = args[0]
        slicer = Slicer.new(self)
        conditions = []
        condition_pairs.each do |key, value|
          case value
          when Range
            # TODO: Optimize "begin <= key <= end" case by missing "between" kernel
            # https://issues.apache.org/jira/browse/ARROW-9843
            unless value.begin.nil?
              conditions << (slicer[key] >= value.begin)
            end
            unless value.end.nil?
              if value.exclude_end?
                conditions << (slicer[key] < value.end)
              else
                conditions << (slicer[key] <= value.end)
              end
            end
          else
            conditions << (slicer[key] == value)
          end
        end
        slicers << conditions.inject(:&)
      else
        slicers << args[0]
      end
    when 2
      offset, length = args
      slicers << (offset...(offset + length))
    else
      expected_n_args = "1..2"
    end
    if expected_n_args
      message = "wrong number of arguments " +
        "(given #{args.size}, expected #{expected_n_args})"
      raise ArgumentError, message
    end
  end

  sliced_tables = []
  slicers.each do |slicer|
    slicer = slicer.evaluate if slicer.respond_to?(:evaluate)
    case slicer
    when Integer
      slicer += n_rows if slicer < 0
      sliced_tables << slice_by_range(slicer, n_rows - 1)
    when Range
      original_from = from = slicer.first
      to = slicer.last
      to -= 1 if slicer.exclude_end?
      from += n_rows if from < 0
      if from < 0 or from >= n_rows
        message =
          "offset is out of range (-#{n_rows + 1},#{n_rows}): " +
          "#{original_from}"
        raise ArgumentError, message
      end
      to += n_rows if to < 0
      sliced_tables << slice_by_range(from, to)
    when ::Array, BooleanArray, ChunkedArray
      sliced_tables << filter(slicer)
    else
      message = "slicer must be Integer, Range, (from, to), " +
        "Arrow::ChunkedArray of Arrow::BooleanArray, " +
        "Arrow::BooleanArray or Arrow::Slicer::Condition: #{slicer.inspect}"
      raise ArgumentError, message
    end
  end
  if sliced_tables.size > 1
    sliced_table = sliced_tables[0].concatenate(sliced_tables[1..-1])
  else
    sliced_table = sliced_tables[0]
  end
  share_input(sliced_table)
  sliced_table
end

#slice_raw ⇒ Object 



200
 
# File 'lib/arrow/table.rb', line 200

alias_method :slice_raw, :slice

#to_s(options = {}) ⇒ Object 



595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
 
# File 'lib/arrow/table.rb', line 595

def to_s(options={})
  format = options[:format]
  case format
  when :column
    return to_s_raw
  when :list
    formatter_class = TableListFormatter
  when :table, nil
    formatter_class = TableTableFormatter
  else
    message = ":format must be :column, :list, :table or nil"
    raise ArgumentError, "#{message}: <#{format.inspect}>"
  end
  formatter = formatter_class.new(self, options)
  formatter.format
end

#to_s_raw ⇒ Object 



594
 
# File 'lib/arrow/table.rb', line 594

alias_method :to_s_raw, :to_s

#window(size: nil) ⇒ Object

Experimental



441
442
443
 
# File 'lib/arrow/table.rb', line 441

def window(size: nil)
  RollingWindow.new(self, size)
end