Class: Arrow::Table

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

Class Method Summary collapse

Instance Method Summary collapse

Methods included from RecordContainable

#each_record

Methods included from GenericTakeable

included, #take_generic

Methods included from GenericFilterable

#filter_generic, included

Methods included from ColumnContainable

#[], #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.

    • arrays (::Array<::Array>) —

      The data of the table as primitive Ruby objects.



162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
 
# File 'lib/arrow/table.rb', line 162

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 



479
480
481
482
483
484
485
 
# File 'lib/arrow/table.rb', line 479

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 



28
29
30
 
# File 'lib/arrow/table.rb', line 28

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

Instance Method Details

#each_record_batch ⇒ Object 



186
187
188
189
190
191
192
193
 
# File 'lib/arrow/table.rb', line 186

def each_record_batch
  return to_enum(__method__) unless block_given?

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

#group(*keys) ⇒ Object

Experimental



430
431
432
 
# File 'lib/arrow/table.rb', line 430

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

#inspect ⇒ Object 



470
471
472
 
# File 'lib/arrow/table.rb', line 470

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

#inspect_raw ⇒ Object 



469
 
# File 'lib/arrow/table.rb', line 469

alias_method :inspect_raw, :inspect

#merge(other) ⇒ Arrow::Table

TODO

Returns:



358
359
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
 
# File 'lib/arrow/table.rb', line 358

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
  self.class.new(new_fields, new_arrays)
end

#pack ⇒ Object 



444
445
446
447
448
449
 
# File 'lib/arrow/table.rb', line 444

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

#remove_column(name_or_index) ⇒ Object 



408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
 
# File 'lib/arrow/table.rb', line 408

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
  remove_column_raw(index)
end

#remove_column_raw ⇒ Object 



407
 
# File 'lib/arrow/table.rb', line 407

alias_method :remove_column_raw, :remove_column

#respond_to_missing?(name, include_private) ⇒ Boolean

Returns:

  • (Boolean) 


474
475
476
477
 
# File 'lib/arrow/table.rb', line 474

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

#save(output, options = {}) ⇒ Object 



439
440
441
442
 
# File 'lib/arrow/table.rb', line 439

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.



253
254
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
 
# File 'lib/arrow/table.rb', line 253

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

  filter_options = Arrow::FilterOptions.new
  filter_options.null_selection_behavior = :emit_null
  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, filter_options)
    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_tables[0].concatenate(sliced_tables[1..-1])
  else
    sliced_tables[0]
  end
end

#slice_raw ⇒ Object 



198
 
# File 'lib/arrow/table.rb', line 198

alias_method :slice_raw, :slice

#to_s(options = {}) ⇒ Object 



452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
 
# File 'lib/arrow/table.rb', line 452

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 



451
 
# File 'lib/arrow/table.rb', line 451

alias_method :to_s_raw, :to_s

#window(size: nil) ⇒ Object

Experimental



435
436
437
 
# File 'lib/arrow/table.rb', line 435

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