Github user DazhuangSu commented on a diff in the pull request:
https://github.com/apache/spark/pull/19691#discussion_r191823358
--- Diff:
sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/parser/AstBuilder.scala
---
@@ -293,6 +313,15 @@ class AstBuilder(conf: SQLConf) extends
SqlBaseBaseVisitor[AnyRef] with Logging
}
}
+ /**
+ * Create a partition specification map without optional values
+ * and a partition filter specification.
+ */
+ protected def visitPartition(
--- End diff --
I tried to add a new parameter to `AlterTableDropPartitionCommand` earlier,
but it was kind hard.
thinking about a sql below:
`DROP PARTITION(partitionVal1, expression1), PARTITION(partitionVal2,
expression2)`
all of the partitions need to be dropped are:
(`partitionVal1` <b>intersect</b> `expression1`) <b>union</b>
(`partitionVal2` <b>intersect</b> `expression2`)
using one tuple is to telling us that the `partitionVal1` and `expression1`
are from the same `partitionSpec` and we should use `intersect`.
Also, different tuples means (`partitionVal1 intersect expression1`) and
(`partitionVal2 intersect expression2`) are from different `partitionSpec` and
we should use `union`.
if we don't use tuple, it's would be difficult to tell the different
occasions and it would be difficult to decide between `intersect` and `union`
when `partitionVal1` meet `expression1`/`expression2`
Any ideas to replace this `tuple`?
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
