Fix Hudi to not write in Parquet legacy format
Open
from-jira
priority:high
type:bug
- Dominant language
- Java
- Stars
- 6.2k
- Forks
- 2.5k
- Avg merge
- 2d 8h
- Merged PRs (30d)
- 111
Description
Currently Hudi have to write in Parquet legacy-format ("spark.sql.parquet.writeLegacyFormat") whenever schema contains Decimals, due to the fact that it relies on AvroParquetReader which is unable to read Decimals in the non-legacy format (ie it could only read Decimals encoded as FIXED_BYTE_ARRAY, and not as INT32/INT64)
This leads to suboptimal storage footprint where for example on some datasets this could lead to a bloat of 10% or more.
## JIRA info
- Link: https://issues.apache.org/jira/browse/HUDI-4321
- Type: Bug
- Epic: https://issues.apache.org/jira/browse/HUDI-3217
- Fix version(s):
- 1.1.0
Contributor guide
No contributing guide indexed for this repository
Assessment
This issue has not been assessed yet.