Datasets:

unkadlabs
/

qor-af-soomaali-frequency

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 2 new columns ({'word', 'sentence_count'}) and 1 missing columns ({'bigram'}).

This happened while the csv dataset builder was generating data using

hf://datasets/unkadlabs/qor-af-soomaali-frequency/data/unigrams.tsv (at revision d38bbb65f21f09b00aa6cd97b10a7b360b2c7251), ['hf://datasets/unkadlabs/qor-af-soomaali-frequency@d38bbb65f21f09b00aa6cd97b10a7b360b2c7251/data/bigrams.tsv', 'hf://datasets/unkadlabs/qor-af-soomaali-frequency@d38bbb65f21f09b00aa6cd97b10a7b360b2c7251/data/unigrams.tsv']

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
              rank: int64
              word: string
              count: int64
              sentence_count: int64
              -- schema metadata --
              pandas: '{"index_columns": [{"kind": "range", "name": null, "start": 0, "' + 725
              to
              {'rank': Value('int64'), 'bigram': Value('string'), 'count': Value('int64')}
              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 2 new columns ({'word', 'sentence_count'}) and 1 missing columns ({'bigram'}).
              
              This happened while the csv dataset builder was generating data using
              
              hf://datasets/unkadlabs/qor-af-soomaali-frequency/data/unigrams.tsv (at revision d38bbb65f21f09b00aa6cd97b10a7b360b2c7251), ['hf://datasets/unkadlabs/qor-af-soomaali-frequency@d38bbb65f21f09b00aa6cd97b10a7b360b2c7251/data/bigrams.tsv', 'hf://datasets/unkadlabs/qor-af-soomaali-frequency@d38bbb65f21f09b00aa6cd97b10a7b360b2c7251/data/unigrams.tsv']
              
              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.

rank int64	bigram string	count int64
1	ah oo	80
2	ma aha	60
3	ka mid	59
4	oo aan	56
5	mid ah	53
6	kasta oo	49
7	waxa ay	46
8	sidoo kale	46
9	waxa uu	46
10	taas oo	44
11	ay ku	44
12	waa in	42
13	ah ee	41
14	oo ka	40
15	waa mid	36
16	in la	36
17	oo ay	35
18	ay ka	34
19	waa inuu	33
20	wuxuu ku	32
21	oo ku	32
22	in ay	32
23	uu ku	31
24	aan la	29
25	ay tahay	29
26	waxaa lagu	29
27	mid ka	29
28	oo lagu	29
29	oo keliya	28
30	oo ah	27
31	kale oo	27
32	badan oo	27
33	ah in	26
34	gaar ahaan	26
35	waxaa la	25
36	oo la	24
37	gaar ah	24
38	wax walba	24
39	marka aad	24
40	marka la	23
41	in aad	23
42	ayaa ka	23
43	ka badan	23
44	ka hor	22
45	waxaa laga	22
46	laga yaabaa	22
47	ugu weyn	22
48	cusub oo	22
49	oo dhan	22
50	waxay ka	22
51	casriga ah	22
52	oo aad	21
53	sidaas darteed	21
54	waxay ku	21
55	ku nool	20
56	ku dhisan	20
57	si ay	20
58	kala duwan	20
59	ku soo	20
60	aan ku	19
61	si uu	19
62	in uu	19
63	dhab ah	19
64	uu ka	18
65	isla markaana	18
66	marka uu	18
67	mararka qaar	18
68	uu yahay	18
69	in aan	18
70	oo uu	18
71	ka soo	17
72	ee uu	17
73	weyn oo	16
74	oo si	16
75	ku saabsan	16
76	waxa aan	16
77	dadka kale	16
78	sababtoo ah	16
79	wax ka	16
80	inta badan	16
81	balse ay	16
82	ah ayaa	16
83	ayaa lagu	16
84	hoggaamiyuhu waa	16
85	si cad	16
86	kuwaas oo	16
87	si loo	16
88	waxaa loo	16
89	la soo	15
90	qurux badan	15
91	qof kasta	15
92	ee aan	15
93	haddii aad	15
94	wax kasta	14
95	joogto ah	14
96	oo kale	14
97	mar walba	14
98	ay yihiin	14
99	wuxuu ka	14
100	badan ee	14

End of preview.

Qor Somali Frequency List (from corpus v0.2.1)

Word and bigram frequencies computed from unkadlabs/qor-af-soomaali v0.2.1: 2,282 sentences of Somali, each written by a consenting, credited Somali speaker, accepted by two independent community validators, and signed off by a linguist reviewer on qor.unkad.com.

To our knowledge this is the first Somali frequency list in which the origin of every counted occurrence is known. Frequency lists derived from web crawls inherit whatever the crawl contains: machine-translated pages, boilerplate, spam, and undetected English. This one inherits only what named people wrote and other people verified.

Files

data/unigrams.tsv — rank, word, count, sentence_count (8,971 words, 36,840 tokens)
data/bigrams.tsv — rank, bigram, count (all bigrams occurring twice or more)
data/stats.json — version lock and totals

Method

Unicode NFC, lowercased. A token is a run of letters with internal apostrophes or hyphens kept (la'aan, af-soomaali); punctuation stripped; numbers and single characters dropped. Nothing else is filtered: the corpus contains occasional code-switched English, and a frequency list that silently edits the language misrepresents it. The list is regenerated per corpus release and version-locked; this one is v0.2.1.

Honest limitations

2,282 sentences is small. Content words in the tail have counts of 1-3 and their ranks are not stable; function-word ranks are already stable. The corpus skews toward the sectors its contributors chose to write about (see the corpus card). The list will be regenerated as the corpus grows toward its 100,000-sentence goal.

Uses we had in mind

Keyboard and autocomplete word lists, spell-checker dictionaries, teaching materials, and a clean reference against which crawled Somali resources can be checked for contamination.

Licence and credit

CC BY-SA 4.0, inherited from the corpus. The people who wrote the underlying sentences are credited by name (where they chose) in the corpus release's CREDITS file. If you use this list, credit "Unkad Labs and the Qor Af-Soomaali contributors".

Downloads last month: 31

Total file size:

220 kB