Datasets:

wkaandemir
/

biostat-survival

Error code:   DatasetGenerationCastError
Exception:    DatasetGenerationCastError
Message:      An error occurred while generating the dataset

All the data files must have the same columns, but at some point there are 8 new columns ({'wexp', 'fin', 'arrest', 'prio', 'week', 'mar', 'paro', 'race'}) and 9 missing columns ({'ph.ecog', 'time', 'sex', 'inst', 'ph.karno', 'status', 'meal.cal', 'wt.loss', 'pat.karno'}).

This happened while the csv dataset builder was generating data using

hf://datasets/wkaandemir/biostat-survival/rossi_recidivism.csv (at revision 00c80d3321ed52f8bb33258a8c91ef7441da56ed), ['hf://datasets/wkaandemir/biostat-survival@00c80d3321ed52f8bb33258a8c91ef7441da56ed/lung_cancer.csv', 'hf://datasets/wkaandemir/biostat-survival@00c80d3321ed52f8bb33258a8c91ef7441da56ed/rossi_recidivism.csv']

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)
Traceback:    Traceback (most recent call last):
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1837, in _prepare_split_single
                  writer.write_table(table)
                  ~~~~~~~~~~~~~~~~~~^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/arrow_writer.py", line 765, in write_table
                  self._write_table(pa_table, writer_batch_size=writer_batch_size)
                  ~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/arrow_writer.py", line 773, in _write_table
                  pa_table = table_cast(pa_table, self._schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2369, in table_cast
                  return cast_table_to_schema(table, schema)
                File "/usr/local/lib/python3.14/site-packages/datasets/table.py", line 2297, in cast_table_to_schema
                  raise CastError(
                  ...<3 lines>...
                  )
              datasets.table.CastError: Couldn't cast
              week: int64
              arrest: int64
              fin: int64
              age: int64
              race: int64
              wexp: int64
              mar: int64
              paro: int64
              prio: int64
              -- schema metadata --
              pandas: '{"index_columns": [{"kind": "range", "name": null, "start": 0, "' + 1222
              to
              {'inst': Value('float64'), 'time': Value('int64'), 'status': Value('int64'), 'age': Value('int64'), 'sex': Value('int64'), 'ph.ecog': Value('float64'), 'ph.karno': Value('float64'), 'pat.karno': Value('float64'), 'meal.cal': Value('float64'), 'wt.loss': Value('float64')}
              because column names don't match
              
              During handling of the above exception, another exception occurred:
              
              Traceback (most recent call last):
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 1369, in compute_config_parquet_and_info_response
                  parquet_operations, partial, estimated_dataset_info = stream_convert_to_parquet(
                                                                        ~~~~~~~~~~~~~~~~~~~~~~~~~^
                      builder, max_dataset_size_bytes=max_dataset_size_bytes
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  )
                  ^
                File "/src/services/worker/src/worker/job_runners/config/parquet_and_info.py", line 948, in stream_convert_to_parquet
                  builder._prepare_split(split_generator=splits_generators[split], file_format="parquet")
                  ~~~~~~~~~~~~~~~~~~~~~~^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1683, in _prepare_split
                  for job_id, done, content in self._prepare_split_single(
                                               ~~~~~~~~~~~~~~~~~~~~~~~~~~^
                      gen_kwargs=gen_kwargs, job_id=job_id, **_prepare_split_args
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                  ):
                  ^
                File "/usr/local/lib/python3.14/site-packages/datasets/builder.py", line 1839, in _prepare_split_single
                  raise DatasetGenerationCastError.from_cast_error(
                  ...<4 lines>...
                  )
              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 8 new columns ({'wexp', 'fin', 'arrest', 'prio', 'week', 'mar', 'paro', 'race'}) and 9 missing columns ({'ph.ecog', 'time', 'sex', 'inst', 'ph.karno', 'status', 'meal.cal', 'wt.loss', 'pat.karno'}).
              
              This happened while the csv dataset builder was generating data using
              
              hf://datasets/wkaandemir/biostat-survival/rossi_recidivism.csv (at revision 00c80d3321ed52f8bb33258a8c91ef7441da56ed), ['hf://datasets/wkaandemir/biostat-survival@00c80d3321ed52f8bb33258a8c91ef7441da56ed/lung_cancer.csv', 'hf://datasets/wkaandemir/biostat-survival@00c80d3321ed52f8bb33258a8c91ef7441da56ed/rossi_recidivism.csv']
              
              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)

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

inst float64	time int64	status int64	age int64	sex int64	ph.ecog float64	ph.karno float64	pat.karno float64	meal.cal float64	wt.loss float64
3	306	1	74	1	1	90	100	1,175	null
3	455	1	68	1	0	90	90	1,225	15
3	1,010	0	56	1	0	90	90	null	15
5	210	1	57	1	1	90	60	1,150	11
1	883	1	60	1	0	100	90	null	0
12	1,022	0	74	1	1	50	80	513	0
7	310	1	68	2	2	70	60	384	10
11	361	1	71	2	2	60	80	538	1
1	218	1	53	1	1	70	80	825	16
7	166	1	61	1	2	70	70	271	34
6	170	1	57	1	1	80	80	1,025	27
16	654	1	68	2	2	70	70	null	23
11	728	1	68	2	1	90	90	null	5
21	71	1	60	1	null	60	70	1,225	32
12	567	1	57	1	1	80	70	2,600	60
1	144	1	67	1	1	80	90	null	15
22	613	1	70	1	1	90	100	1,150	-5
16	707	1	63	1	2	50	70	1,025	22
1	61	1	56	2	2	60	60	238	10
21	88	1	57	1	1	90	80	1,175	null
11	301	1	67	1	1	80	80	1,025	17
6	81	1	49	2	0	100	70	1,175	-8
11	624	1	50	1	1	70	80	null	16
15	371	1	58	1	0	90	100	975	13
12	394	1	72	1	0	90	80	null	0
12	520	1	70	2	1	90	80	825	6
4	574	1	60	1	0	100	100	1,025	-13
13	118	1	70	1	3	60	70	1,075	20
13	390	1	53	1	1	80	70	875	-7
1	12	1	74	1	2	70	50	305	20
12	473	1	69	2	1	90	90	1,025	-1
1	26	1	73	1	2	60	70	388	20
7	533	1	48	1	2	60	80	null	-11
16	107	1	60	2	2	50	60	925	-15
12	53	1	61	1	2	70	100	1,075	10
1	122	1	62	2	2	50	50	1,025	null
22	814	1	65	1	2	70	60	513	28
15	965	0	66	2	1	70	90	875	4
1	93	1	74	1	2	50	40	1,225	24
1	731	1	64	2	1	80	100	1,175	15
5	460	1	70	1	1	80	60	975	10
11	153	1	73	2	2	60	70	1,075	11
10	433	1	59	2	0	90	90	363	27
12	145	1	60	2	2	70	60	null	null
7	583	1	68	1	1	60	70	1,025	7
7	95	1	76	2	2	60	60	625	-24
1	303	1	74	1	0	90	70	463	30
3	519	1	63	1	1	80	70	1,025	10
13	643	1	74	1	0	90	90	1,425	2
22	765	1	50	2	1	90	100	1,175	4
3	735	1	72	2	1	90	90	null	9
12	189	1	63	1	0	80	70	null	0
21	53	1	68	1	0	90	100	1,025	0
1	246	1	58	1	0	100	90	1,175	7
6	689	1	59	1	1	90	80	1,300	15
1	65	1	62	1	0	90	80	725	null
5	5	1	65	2	0	100	80	338	5
22	132	1	57	1	2	70	60	null	18
3	687	1	58	2	1	80	80	1,225	10
1	345	1	64	2	1	90	80	1,075	-3
22	444	1	75	2	2	70	70	438	8
12	223	1	48	1	1	90	80	1,300	68
21	175	1	73	1	1	80	100	1,025	null
11	60	1	65	2	1	90	80	1,025	0
3	163	1	69	1	1	80	60	1,125	0
3	65	1	68	1	2	70	50	825	8
16	208	1	67	2	2	70	null	538	2
5	821	0	64	2	0	90	70	1,025	3
22	428	1	68	1	0	100	80	1,039	0
6	230	1	67	1	1	80	100	488	23
13	840	0	63	1	0	90	90	1,175	-1
3	305	1	48	2	1	80	90	538	29
5	11	1	74	1	2	70	100	1,175	0
2	132	1	40	1	1	80	80	null	3
21	226	1	53	2	1	90	80	825	3
12	426	1	71	2	1	90	90	1,075	19
1	705	1	51	2	0	100	80	1,300	0
6	363	1	56	2	1	80	70	1,225	-2
3	11	1	81	1	0	90	null	731	15
1	176	1	73	1	0	90	70	169	30
4	791	1	59	1	0	100	80	768	5
13	95	1	55	1	1	70	90	1,500	15
11	196	0	42	1	1	80	80	1,425	8
21	167	1	44	2	1	80	90	588	-1
16	806	0	44	1	1	80	80	1,025	1
6	284	1	71	1	1	80	90	1,100	14
22	641	1	62	2	1	80	80	1,150	1
21	147	1	61	1	0	100	90	1,175	4
13	740	0	44	2	1	90	80	588	39
1	163	1	72	1	2	70	70	910	2
11	655	1	63	1	0	100	90	975	-1
22	239	1	70	1	1	80	100	null	23
5	88	1	66	1	1	90	80	875	8
10	245	1	57	2	1	80	60	280	14
1	588	0	69	2	0	100	90	null	13
12	30	1	72	1	2	80	60	288	7
3	179	1	69	1	1	80	80	null	25
12	310	1	71	1	1	90	100	null	0
11	477	1	64	1	1	90	100	910	0
3	166	1	70	2	0	90	70	null	10

End of preview.

Biostatistics Survival Analysis Datasets

Two classic time-to-event datasets for survival analysis benchmarking.

Datasets

1. Rossi Recidivism (rossi_recidivism.csv)

Recidivism study of 432 former prisoners released from Maryland state prisons. Duration measured in weeks; event is arrest.

Rows: 432
Duration column: week
Event column: arrest (1=arrested, 0=censored)
Censoring: 73.6%
Source: Rossi et al. (1980). Money, Work, and Crime: Some Experimental Results.

2. NCCTG Lung Cancer (lung_cancer.csv)

North Central Cancer Treatment Group advanced lung cancer survival data. Duration measured in days; event is death.

Rows: 228 (167 complete cases)
Duration column: time
Event column: status (2=dead, 1=alive/censored)
Censoring: 27.6%
Source: Loprinzi et al. (1994). Prospective evaluation of prognostic variables. J Clin Oncol.

Benchmark Results

Cox Proportional Hazards Model

Dataset	C-Index	Logrank p	Median Survival
rossi_recidivism	0.6403	0.0501	inf
lung_cancer	0.6532	0.0139	310.0

Key Hazard Ratios (Cox PH)

Rossi Recidivism

fin (financial aid): HR=0.684, p=0.047 -- financial aid reduces recidivism risk
age: HR=0.944, p=0.009 -- older prisoners less likely to reoffend
prio (prior arrests): HR=1.096, p=0.001 -- more prior arrests increases risk

Lung Cancer

sex: HR=0.575, p=0.006 -- females have lower mortality risk
ph.ecog (ECOG performance score): HR=2.095, p=0.001 -- stronger predictor
ph.karno (Karnofsky score): HR=1.023, p=0.046

Methods

Kaplan-Meier estimator for survival curves
Logrank test for group comparisons
Cox Proportional Hazards regression for hazard ratios
5-fold cross-validation for C-index estimation

Citation

See source references above. These datasets are repackaged for HF datasets format.

Downloads last month: 16

Total file size:

22.5 kB