All the data files must have the same columns, but at some point there are 3 new columns ({'source', 'version', 'metadata'})
#2
by
davidlau90
- opened
Hi, i've encountered a problem when load_dataest():
dataset = load_dataset("allenai/olmo-mix-1124", split="train", trust_remote_code=True, verification_mode=datasets.VerificationMode.NO_CHECKS)
Resolving data files: 100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 28/28 [00:00<00:00, 247243.18it/s]
Resolving data files: 100%|█████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 28/28 [00:00<00:00, 224123.11it/s]
Downloading data: 100%|██████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████████| 28/28 [00:00<00:00, 32.52files/s]
Generating train split: 0 examples [00:00, ? examples/s]
Traceback (most recent call last):
File "/home//.local/lib/python3.10/site-packages/datasets/builder.py", line 1870, in _prepare_split_single
writer.write_table(table)
File "/home//.local/lib/python3.10/site-packages/datasets/arrow_writer.py", line 622, in write_table
pa_table = table_cast(pa_table, self._schema)
File "/home//.local/lib/python3.10/site-packages/datasets/table.py", line 2292, in table_cast
return cast_table_to_schema(table, schema)
File "/home//.local/lib/python3.10/site-packages/datasets/table.py", line 2240, in cast_table_to_schema
raise CastError(
datasets.table.CastError: Couldn't cast
added: string
created: string
id: string
metadata: struct<abstract: string, abstract_count: int64, abstract_language: string, abstract_perplexity: double, extfieldsofstudy: list<item: string>, provenance: string, s2fieldsofstudy: list<item: string>, sha1: string, sources: list<item: string>, title: string, title_count: int64, title_language: string, title_perplexity: double, top_frequencies: list<item: struct<count: int64, token: string>>, year: int64>
child 0, abstract: string
child 1, abstract_count: int64
child 2, abstract_language: string
child 3, abstract_perplexity: double
child 4, extfieldsofstudy: list<item: string>
child 0, item: string
child 5, provenance: string
child 6, s2fieldsofstudy: list<item: string>
child 0, item: string
child 7, sha1: string
child 8, sources: list<item: string>
child 0, item: string
child 9, title: string
child 10, title_count: int64
child 11, title_language: string
child 12, title_perplexity: double
child 13, top_frequencies: list<item: struct<count: int64, token: string>>
child 0, item: struct<count: int64, token: string>
child 0, count: int64
child 1, token: string
child 14, year: int64
source: string
text: string
version: string
to
{'id': Value(dtype='string', id=None), 'text': Value(dtype='string', id=None), 'added': Value(dtype='string', id=None), 'created': Value(dtype='string', id=None)}
because column names don't match
During handling of the above exception, another exception occurred:
Traceback (most recent call last):
File "<stdin>", line 1, in <module>
File "/home//.local/lib/python3.10/site-packages/datasets/load.py", line 2151, in load_dataset
builder_instance.download_and_prepare(
File "/home//.local/lib/python3.10/site-packages/datasets/builder.py", line 924, in download_and_prepare
self._download_and_prepare(
File "/home//.local/lib/python3.10/site-packages/datasets/builder.py", line 1000, in _download_and_prepare
self._prepare_split(split_generator, **prepare_split_kwargs)
File "/home//.local/lib/python3.10/site-packages/datasets/builder.py", line 1741, in _prepare_split
for job_id, done, content in self._prepare_split_single(
File "/home//.local/lib/python3.10/site-packages/datasets/builder.py", line 1872, in _prepare_split_single
raise DatasetGenerationCastError.from_cast_error(
datasets.exceptions.DatasetGenerationCastError: An error occurred while generating the dataset
All the data files must have the same columns, but at some point there are 3 new columns ({'source', 'version', 'metadata'})
This happened while the json dataset builder was generating data using
/home//.cache/huggingface/hub/datasets--allenai--olmo-mix-1124/snapshots/166bbe2db8563a30388e37b4cfd753b8252f3622/data/pes2o/pes2o-0000.json.gz
Please either edit the data files to have matching columns, or separate them into different configurations (see docs at https://hf.co/docs/hub/datasets-manual-configuration#multiple-configurations)
Could you please help?
A workaround (which is not ideal) is to manually define specific data files and features when loading. For example:
from datasets import load_dataset, Features, Value
features = Features({
'text': Value('string'),
'id': Value('string'),
'metadata': Features({
'length': Value('int64'),
'provenance': Value('string'),
'revid': Value('string'),
'url': Value('string')
}),
...
})
dataset = load_dataset("allenai/olmo-mix-1124", data_files="data/wiki/*", features=features)
There are actually a few mistakes in the dataset metadata that cause errors when loading this dataset. For instance, the configuration (in README) specifies:
- config_name: dclm
data_files:
- split: train
path: data/dclm/*
but the actual data files are located at:data/dclm/raw/hero-run-fasttext_for_HF/filtered/OH_eli5_vs_rw_v2_bigram_200k_train/fasttext_openhermes_reddit_eli5_vs_rw_v2_bigram_200k_train/processed_data/global-shard_[[xx]]_of_10/local-shard_[[x]]_of_10
I think that's why the preview fails for this subset, too. Hope they'll fix it🙏🏻
Hi, thanks all for the feedback, and apologies for the inconvenience. We have put in some work to solve this issue and it should now be resolved!