Add column_names to IterableDataset

[patrickloeber]

This PR closes #5383

• Add column_names property to IterableDataset
• Add multiple tests for this new property

[HuggingFaceDocBuilderDev]

The documentation is not available anymore as the PR was closed or merged.

[mariosasko]

Thanks!

The CI failures are unrelated to this PR.

[lhoestq]

thanks !

[github-actions]

Show benchmarks

PyArrow==8.0.0

Show updated benchmarks!

Benchmark: benchmark_array_xd.json

metric	read_batch_formatted_as_numpy after write_array2d	read_batch_formatted_as_numpy after write_flattened_sequence	read_batch_formatted_as_numpy after write_nested_sequence	read_batch_unformated after write_array2d	read_batch_unformated after write_flattened_sequence	read_batch_unformated after write_nested_sequence	read_col_formatted_as_numpy after write_array2d	read_col_formatted_as_numpy after write_flattened_sequence	read_col_formatted_as_numpy after write_nested_sequence	read_col_unformated after write_array2d	read_col_unformated after write_flattened_sequence	read_col_unformated after write_nested_sequence	read_formatted_as_numpy after write_array2d	read_formatted_as_numpy after write_flattened_sequence	read_formatted_as_numpy after write_nested_sequence	read_unformated after write_array2d	read_unformated after write_flattened_sequence	read_unformated after write_nested_sequence	write_array2d	write_flattened_sequence	write_nested_sequence
new / old (diff)	0.006362 / 0.011353 (-0.004991)	0.004546 / 0.011008 (-0.006462)	0.097003 / 0.038508 (0.058495)	0.028007 / 0.023109 (0.004898)	0.315097 / 0.275898 (0.039199)	0.365128 / 0.323480 (0.041649)	0.004819 / 0.007986 (-0.003167)	0.003335 / 0.004328 (-0.000994)	0.076665 / 0.004250 (0.072415)	0.038285 / 0.037052 (0.001233)	0.322100 / 0.258489 (0.063611)	0.407466 / 0.293841 (0.113625)	0.031580 / 0.128546 (-0.096966)	0.011645 / 0.075646 (-0.064001)	0.321789 / 0.419271 (-0.097483)	0.051015 / 0.043533 (0.007483)	0.331762 / 0.255139 (0.076623)	0.369727 / 0.283200 (0.086527)	0.090144 / 0.141683 (-0.051539)	1.485480 / 1.452155 (0.033326)	1.562032 / 1.492716 (0.069316)

Benchmark: benchmark_getitem_100B.json

metric	get_batch_of_1024_random_rows	get_batch_of_1024_rows	get_first_row	get_last_row
new / old (diff)	0.201192 / 0.018006 (0.183186)	0.409760 / 0.000490 (0.409270)	0.002220 / 0.000200 (0.002020)	0.000070 / 0.000054 (0.000016)

Benchmark: benchmark_indices_mapping.json

metric	select	shard	shuffle	sort	train_test_split
new / old (diff)	0.022361 / 0.037411 (-0.015050)	0.096375 / 0.014526 (0.081849)	0.101369 / 0.176557 (-0.075188)	0.161568 / 0.737135 (-0.575568)	0.105094 / 0.296338 (-0.191245)

Benchmark: benchmark_iterating.json

metric	read 5000	read 50000	read_batch 50000 10	read_batch 50000 100	read_batch 50000 1000	read_formatted numpy 5000	read_formatted pandas 5000	read_formatted tensorflow 5000	read_formatted torch 5000	read_formatted_batch numpy 5000 10	read_formatted_batch numpy 5000 1000	shuffled read 5000	shuffled read 50000	shuffled read_batch 50000 10	shuffled read_batch 50000 100	shuffled read_batch 50000 1000	shuffled read_formatted numpy 5000	shuffled read_formatted_batch numpy 5000 10	shuffled read_formatted_batch numpy 5000 1000
new / old (diff)	0.426251 / 0.215209 (0.211042)	4.261374 / 2.077655 (2.183720)	2.015688 / 1.504120 (0.511569)	1.833708 / 1.541195 (0.292513)	1.908994 / 1.468490 (0.440504)	0.703108 / 4.584777 (-3.881669)	3.420767 / 3.745712 (-0.324945)	1.844776 / 5.269862 (-3.425086)	1.158470 / 4.565676 (-3.407207)	0.083324 / 0.424275 (-0.340951)	0.013054 / 0.007607 (0.005447)	0.521473 / 0.226044 (0.295429)	5.245505 / 2.268929 (2.976576)	2.349110 / 55.444624 (-53.095515)	2.011119 / 6.876477 (-4.865358)	2.217807 / 2.142072 (0.075734)	0.808584 / 4.805227 (-3.996643)	0.151337 / 6.500664 (-6.349327)	0.065815 / 0.075469 (-0.009654)

Benchmark: benchmark_map_filter.json

metric	filter	map fast-tokenizer batched	map identity	map identity batched	map no-op batched	map no-op batched numpy	map no-op batched pandas	map no-op batched pytorch	map no-op batched tensorflow
new / old (diff)	1.221839 / 1.841788 (-0.619949)	13.634161 / 8.074308 (5.559853)	13.915360 / 10.191392 (3.723968)	0.126448 / 0.680424 (-0.553976)	0.016614 / 0.534201 (-0.517587)	0.379150 / 0.579283 (-0.200133)	0.382134 / 0.434364 (-0.052230)	0.442845 / 0.540337 (-0.097493)	0.519578 / 1.386936 (-0.867358)

PyArrow==latest

Show updated benchmarks!

Benchmark: benchmark_array_xd.json

metric	read_batch_formatted_as_numpy after write_array2d	read_batch_formatted_as_numpy after write_flattened_sequence	read_batch_formatted_as_numpy after write_nested_sequence	read_batch_unformated after write_array2d	read_batch_unformated after write_flattened_sequence	read_batch_unformated after write_nested_sequence	read_col_formatted_as_numpy after write_array2d	read_col_formatted_as_numpy after write_flattened_sequence	read_col_formatted_as_numpy after write_nested_sequence	read_col_unformated after write_array2d	read_col_unformated after write_flattened_sequence	read_col_unformated after write_nested_sequence	read_formatted_as_numpy after write_array2d	read_formatted_as_numpy after write_flattened_sequence	read_formatted_as_numpy after write_nested_sequence	read_unformated after write_array2d	read_unformated after write_flattened_sequence	read_unformated after write_nested_sequence	write_array2d	write_flattened_sequence	write_nested_sequence
new / old (diff)	0.006238 / 0.011353 (-0.005115)	0.004591 / 0.011008 (-0.006418)	0.076652 / 0.038508 (0.038144)	0.026882 / 0.023109 (0.003773)	0.341948 / 0.275898 (0.066050)	0.375244 / 0.323480 (0.051764)	0.004770 / 0.007986 (-0.003215)	0.004703 / 0.004328 (0.000374)	0.075797 / 0.004250 (0.071547)	0.035001 / 0.037052 (-0.002051)	0.341670 / 0.258489 (0.083181)	0.383028 / 0.293841 (0.089187)	0.031756 / 0.128546 (-0.096791)	0.011714 / 0.075646 (-0.063933)	0.085552 / 0.419271 (-0.333720)	0.047697 / 0.043533 (0.004164)	0.340805 / 0.255139 (0.085666)	0.365478 / 0.283200 (0.082278)	0.093146 / 0.141683 (-0.048537)	1.465100 / 1.452155 (0.012945)	1.552708 / 1.492716 (0.059992)

Benchmark: benchmark_getitem_100B.json

metric	get_batch_of_1024_random_rows	get_batch_of_1024_rows	get_first_row	get_last_row
new / old (diff)	0.209117 / 0.018006 (0.191111)	0.402622 / 0.000490 (0.402132)	0.003940 / 0.000200 (0.003740)	0.000078 / 0.000054 (0.000023)

Benchmark: benchmark_indices_mapping.json

metric	select	shard	shuffle	sort	train_test_split
new / old (diff)	0.026027 / 0.037411 (-0.011385)	0.098346 / 0.014526 (0.083820)	0.107349 / 0.176557 (-0.069207)	0.157846 / 0.737135 (-0.579289)	0.109566 / 0.296338 (-0.186772)

Benchmark: benchmark_iterating.json

metric	read 5000	read 50000	read_batch 50000 10	read_batch 50000 100	read_batch 50000 1000	read_formatted numpy 5000	read_formatted pandas 5000	read_formatted tensorflow 5000	read_formatted torch 5000	read_formatted_batch numpy 5000 10	read_formatted_batch numpy 5000 1000	shuffled read 5000	shuffled read 50000	shuffled read_batch 50000 10	shuffled read_batch 50000 100	shuffled read_batch 50000 1000	shuffled read_formatted numpy 5000	shuffled read_formatted_batch numpy 5000 10	shuffled read_formatted_batch numpy 5000 1000
new / old (diff)	0.445088 / 0.215209 (0.229879)	4.450727 / 2.077655 (2.373072)	2.237798 / 1.504120 (0.733678)	2.026060 / 1.541195 (0.484866)	2.020464 / 1.468490 (0.551974)	0.700155 / 4.584777 (-3.884622)	3.435497 / 3.745712 (-0.310215)	2.851970 / 5.269862 (-2.417891)	1.512689 / 4.565676 (-3.052988)	0.083717 / 0.424275 (-0.340558)	0.012466 / 0.007607 (0.004859)	0.545130 / 0.226044 (0.319085)	5.478228 / 2.268929 (3.209300)	2.554169 / 55.444624 (-52.890456)	2.214703 / 6.876477 (-4.661774)	2.229997 / 2.142072 (0.087925)	0.809851 / 4.805227 (-3.995376)	0.151019 / 6.500664 (-6.349645)	0.066354 / 0.075469 (-0.009115)

Benchmark: benchmark_map_filter.json

metric	filter	map fast-tokenizer batched	map identity	map identity batched	map no-op batched	map no-op batched numpy	map no-op batched pandas	map no-op batched pytorch	map no-op batched tensorflow
new / old (diff)	1.281016 / 1.841788 (-0.560772)	14.071312 / 8.074308 (5.997004)	14.682465 / 10.191392 (4.491073)	0.144197 / 0.680424 (-0.536227)	0.017088 / 0.534201 (-0.517113)	0.379049 / 0.579283 (-0.200234)	0.390713 / 0.434364 (-0.043650)	0.435804 / 0.540337 (-0.104534)	0.518895 / 1.386936 (-0.868041)