[SPARK-10437] Support aggregation expressions in Order By

• Type: Bug
• Status: Resolved
• Priority: Major
• Resolution: Fixed
• Fix Version/s: 1.6.0
• Component/s: SQL
• Assignee: L. C. Hsieh
• Reporter: Harish Butani
• Created: 03/Sep/15 19:50
• Updated: 16/Sep/15 08:08
• Resolved: 15/Sep/15 20:33

Description

Followup on SPARK-6583

The following still fails.

val df = sqlContext.read.json("examples/src/main/resources/people.json")
df.registerTempTable("t")
val df2 = sqlContext.sql("select age, count(*) from t group by age order by count(*)")
df2.show()

StackTrace

Caused by: org.apache.spark.sql.catalyst.errors.package$TreeNodeException: No function to evaluate expression. type: Count, tree: COUNT(1)
	at org.apache.spark.sql.catalyst.expressions.AggregateExpression.eval(aggregates.scala:41)
	at org.apache.spark.sql.catalyst.expressions.RowOrdering.compare(rows.scala:219)

In 1.4 the issue seemed to be BindReferences.bindReference didn't handle this case.
Haven't looked at 1.5 code, but don't see a change to bindReference in this patch.

Issue Links

• relates to: SPARK-6583 Support aggregated function in order by
• links to: [Github] Pull Request #8599 (viirya)