Github user a10y commented on the issue:
https://github.com/apache/spark/pull/20580
As far as we can tell this is an accidental breaking change, as dropping
support for this in vectorized Parquet reader was never called out. We have
Parquet datasets with binary columns with logical type DECIMAL that were
loadable before the change and have since become unloadable, throwing in
`readBinaryBatch`
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
