Github user liancheng commented on the pull request:
https://github.com/apache/spark/pull/5188#issuecomment-86531248
Hey @budde, sorry that I just merged #5141 to fix a regression, and
introduced conflicts to this PR. As @marmbrus suggested, would you mind to open
a PR against master?
Verified that by disabling the new Parquet data source and metastore
Parquet table conversion, both 1.2.0 and 1.3.0 works in this scenario.
After double thinking about the semantics of the metastore / parquet schema
reconcilation process, I think the rules should be:
1. There shouldn't be any fields share the same name in both schemas but
have different data type
2. The reconciled schema should contain exactly those fields in the
metastore schema
3. When reading a Parquet file with extra fields not existed in the
metastore schema, ignore them
4. When reading a Parquet file with less fields than the metastore schema,
add them as nullable fields (just as what this PR does)
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at infrastruct...@apache.org or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscr...@spark.apache.org
For additional commands, e-mail: reviews-h...@spark.apache.org
