Hi,

Is there a way on the Parquet API to ignore certain columns (or nonexistent
columns) to be filtered by filter2.predicate?

There is a bug on https://issues.apache.org/jira/browse/HIVE-11401 that
when a partition column is used on the WHERE statement, then Hive pushes
down the predicate to Parquet, but parquet fails with column was not found
on schema.

I am thinking to remove such columns from the Hive predicate before calling
Parquet, but I was wondering if the Parquet API allows to ignore them
instead.

Thanks,
- Sergio

Reply via email to