File size: 713 Bytes
1be7f91
 
7da72ab
 
1be7f91
 
 
6d24261
 
7da72ab
 
6d24261
7da72ab
 
 
 
 
 
 
 
 
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
# file_id, ocr, title, date, author, page_count, word_count, character_count

echo """
CREATE TABLE presse AS (
SELECT title
     , author
     , LPAD((REGEXP_EXTRACT(date, '1[0-9][0-9][0-9]') || '-01-01'), 10, '0')::DATE AS year
  FROM read_parquet(
    [('https://huggingface.co/datasets/PleIAs/French-PD-Newspapers/resolve/main/gallica_presse_{:d}.parquet').format(n) for n in range(1, 321)])
);
COPY presse TO '$TMPDIR/presse.parquet' (FORMAT 'parquet', COMPRESSION 'GZIP');
""" | duckdb

# isatty
if [ -t 1 ]; then
  echo parquet file output at: $TMPDIR/presse.parquet
  echo "duckdb -csv :memory: \"SELECT * FROM '$TMPDIR/presse.parquet'\""
else
  cat $TMPDIR/presse.parquet
  rm $TMPDIR/presse.parquet
fi