Parquet cannot read string/byte array field

[raffricc]

Parquet cannot read string/byte array field

[raffricc]

I can confirm that it works in python:

import duckdb
con = duckdb.connect()
results = con.execute("SELECT * FROM '/tmp/writer_test368959290/test.parquet'").fetchall()
print(results)

output:
[(5577006791947779410, b'VlBzgbaiCM', 0.8136399609900968, False), (8763219523051806976, b'CpCILpgRwz', 0.8344520379886149, True), (2727609509146539891, b'peRpmICkHP', 0.28159210753652814, False), (3808190704329644902, b'HUkjEoSJKl', 0.20036864355101555, True)]

[marcboeker]

@raffricc I've added a DUCKDB_TYPE_BLOB that should fix the problem. Please get the latest version and try again. You should see the output as bytes.

2022/06/22 23:39:04 /tmp/writer_test2898174709/test.parquet
2022/06/22 23:39:04 5577006791947779410, VlBzgbaiCM, 0.8136399609900968, true
2022/06/22 23:39:04 2669909356377979369, cWUPuTHNsG, 0.2585587212345129, true
2022/06/22 23:39:04 6828643170743844117, KrTldMuFkY, 0.9464975830653858, false
2022/06/22 23:39:04 8602951731698254683, baxkcGJBFz, 0.4186347320545146, true
2022/06/22 23:39:04 *********************************************
2022/06/22 23:39:04 5577006791947779410, [86 108 66 122 103 98 97 105 67 77], 0.8136399609900968, true

[raffricc]

@marcboeker Hi Marc, thank you very much for the quick fix. Really appreciated. I can confirm it works :)

[marcboeker]

@raffricc Great to hear that it works now :) Thanks for the bug report.

[raffricc]

@marcboeker Morning Marc, I just did some more testing and found a new bug (see #27). Also you might be able to close: #23 by suggesting updating to duckdb 0.4.0