linhongliu-db commented on a change in pull request #34254:
URL: https://github.com/apache/spark/pull/34254#discussion_r727672882
##########
File path:
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/catalog/SessionCatalog.scala
##########
@@ -898,14 +905,24 @@ class SessionCatalog(
val nameToCurrentOrdinal = scala.collection.mutable.HashMap.empty[String,
Int]
val viewDDL = buildViewDDL(metadata, isTempView)
- val projectList = viewColumnNames.zip(metadata.schema).map { case (name,
field) =>
- val normalizedName = normalizeColName(name)
- val count = nameToCounts(normalizedName)
- val ordinal = nameToCurrentOrdinal.getOrElse(normalizedName, 0)
- nameToCurrentOrdinal(normalizedName) = ordinal + 1
- val col = GetViewColumnByNameAndOrdinal(
- metadata.identifier.toString, name, ordinal, count, viewDDL)
- Alias(UpCast(col, field.dataType), field.name)(explicitMetadata =
Some(field.metadata))
+ val projectList = if (!isHiveCreatedView(metadata)) {
+ viewColumnNames.zip(metadata.schema).map { case (name, field) =>
+ val normalizedName = normalizeColName(name)
+ val count = nameToCounts(normalizedName)
+ val ordinal = nameToCurrentOrdinal.getOrElse(normalizedName, 0)
+ nameToCurrentOrdinal(normalizedName) = ordinal + 1
+ val col = GetViewColumnByNameAndOrdinal(
+ metadata.identifier.toString, name, ordinal, count, viewDDL)
+ Alias(UpCast(col, field.dataType), field.name)(explicitMetadata =
Some(field.metadata))
+ }
+ } else {
+ // For view created by hive, the parsed view plan may have different
output columns with
+ // the schema stored in metadata. For example: `CREATE VIEW v AS SELECT
1 FROM t`
+ // the schema in metadata will be `_c0` while the parsed view plan has
column named `1`
+ metadata.schema.zipWithIndex.map { case (field, index) =>
+ val col = GetColumnByOrdinal(index, field.dataType)
Review comment:
`UpCast` is used to handle the data type change when the underlying
table column is altered. In this case, most time, we need to use
`GetViewColumnByNameAndOrdinal` as well.
But for the hive, since we can only use `GetColumnByOrdinal`, so the
`UpCast` couldn't bring many benefits. And if the underlying table is not
changed, we don't need to cast the data type.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
