houqp commented on pull request #792:
URL: https://github.com/apache/arrow-datafusion/pull/792#issuecomment-918380666
OK, so now we are running into a limitation in datafusion. Datafusion
doesn't support filtering on unprojected columns yet. The rule here optimized
the compound expressions into a single dummy projected column, which removes
the individual column projection. I bet if you change it to the following it
would work:
```python
df = df.select(
f.col("a"),
f.col("a") + f.col("b"),
f.col("a") - f.col("b"),
).filter(f.col("a") > f.lit(2))
```
This means the impact of this change is bigger than we expected and could
break queries that are working today. The design and implementation of this
change is correct, it's just it exposed a limitation of datafusion to broader
range of queries :(
--
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]
