RyanLin5967 opened a new issue, #4191:
URL: https://github.com/apache/iceberg-python/issues/4191
### Apache Iceberg version
0.12.0 (latest release)
### Please describe the bug 🐞
Currently, `add_files` takes an identity partition from the column's min and
max and ignores its null count, so a file holding `'a'` and NULL is registered
as partition `'a'` and `IS NULL` filters prune it. A file holding `'a'` and
`'b'` is refused.
```python
import tempfile
import pyarrow as pa, pyarrow.parquet as pq
from pyiceberg.catalog.sql import SqlCatalog
d = tempfile.mkdtemp()
cat = SqlCatalog("c", uri="sqlite:///:memory:", warehouse=f"file://{d}")
cat.create_namespace("ns")
df = pa.table({"x": [1, 2], "s": ["a", None]})
t = cat.create_table("ns.t", schema=df.schema)
with t.update_spec() as u:
u.add_identity("s")
pq.write_table(df, f"{d}/a.parquet")
t.add_files([f"{d}/a.parquet"])
print(t.scan(row_filter="s IS NULL").to_arrow()["x"].to_pylist())
# [], should be [2]
pq.write_table(pa.table({"x": [3, 4], "s": ["a", "b"]}), f"{d}/b.parquet")
t.add_files([f"{d}/b.parquet"])
# ValueError: Cannot infer partition value from parquet metadata as there
are more than one partition values ...
```
`day`, `truncate` and a NaN beside 1.0 do the same, and with
`write.metadata.metrics.default` set to `none` every added file gets a NULL
partition.
tested on pyiceberg 0.12.0 and main
### Willingness to contribute
- [ ] I can contribute a fix for this bug independently
- [ ] I would be willing to contribute a fix for this bug with guidance from
the Iceberg community
- [ ] I cannot contribute a fix for this bug at this time
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
