Github user gatorsmile commented on a diff in the pull request:
https://github.com/apache/spark/pull/18000#discussion_r116805363
--- Diff:
sql/core/src/main/scala/org/apache/spark/sql/execution/datasources/parquet/ParquetFilters.scala
---
@@ -168,6 +168,11 @@ private[parquet] object ParquetFilters {
def createFilter(schema: StructType, predicate: sources.Filter):
Option[FilterPredicate] = {
val dataTypeOf = getFieldMap(schema)
+ // Note that, we do not push down filters for columns having dots in
the names. In Parquet
+ // 1.8.2, currently, column APIs in Parquet's `FilterApi` only allow
dot-separated names
+ // so here we simply avoid this case. See SPARK-20364.
--- End diff --
Parquet does not allow . in the column name because it is used as a column
path delimiter. Since Parquet 1.8.2 (PARQUET-389), Parquet accepts the filter
predicates with missing columns. Since it can generate the incorrect results,
we block the filter pushdown when the filters have the column having dots in
the names. See SPARK-20364.
---
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
