ajantha-bhat commented on a change in pull request #3757:
URL: https://github.com/apache/iceberg/pull/3757#discussion_r771769167
##########
File path:
spark/v3.2/spark/src/main/scala/org/apache/spark/sql/execution/datasources/SparkExpressionConverter.scala
##########
@@ -30,4 +34,18 @@ object SparkExpressionConverter {
// But these two conversions already exist and well tested. So, we are
going with this approach.
SparkFilters.convert(DataSourceStrategy.translateFilter(sparkExpression,
supportNestedPredicatePushdown = true).get)
}
+
+ @throws[AnalysisException]
+ def collectResolvedSparkExpression(session: SparkSession, tableName: String,
where: String): Expression = {
+ var expression:Expression = null
+ // Add a dummy prefix linking to the table to collect the resolved spark
expression from analyzed plan.
+ val prefix = String.format("SELECT 42 from %s where ", tableName)
+ val logicalPlan = session.sessionState.sqlParser.parsePlan(prefix + where)
+ val analyzedLogicalPlan = session.sessionState.executePlan(logicalPlan,
CommandExecutionMode.ALL).analyzed
+ analyzedLogicalPlan.collectFirst {
+ case filter: Filter =>
+ expression = filter.expressions.head
+ }
+ expression
Review comment:
a. we can't use SparkFilters.convert, as Filter in plan is
`org.apache.spark.sql.catalyst.plans.logical.Filter` and Filter in
SparkFilters.convert is `org.apache.spark.sql.sources.Filter`
b. I think we can use collectFirst as the plan is fixed always as query is
fixed and it has only one filter Node. It is simple Project + Filter + scan
operators in the pan.
--
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]
