Github user rxin commented on a diff in the pull request:
https://github.com/apache/spark/pull/4592#discussion_r24731783
--- Diff: sql/core/src/main/scala/org/apache/spark/sql/GroupedData.scala ---
@@ -39,13 +39,14 @@ class GroupedData protected[sql](df: DataFrameImpl,
groupingExprs: Seq[Expressio
df.sqlContext, Aggregate(groupingExprs, namedGroupingExprs ++
aggExprs, df.logicalPlan))
}
- private[this] def aggregateNumericColumns(f: Expression => Expression):
Seq[NamedExpression] = {
- df.numericColumns.map { c =>
- val a = f(c)
- Alias(a, a.toString)()
- }
+ private[this] def aggregateNumericColumns(colNames: String*)
--- End diff --
can you change this to
```scala
private[this] def aggregateNumericColumns(colNames: String*)(f:
Expression => Expression)
: Seq[NamedExpression] = {
val columnExprs = if (colNames.isEmpty) {
// No columns specified. Use all numeric columns.
df.numericColumns
} else {
// Make sure all specified columns are numeric
colNames.map { colName =>
val namedExpr = df.resolve(colName)
if (!namedExpr.dataType.isInstanceOf[NumericType]) {
throw new AnalysisException(
s""""$colName" is not a numeric column. """ +
"Aggregation function can only be performed on a numeric
column.")
}
namedExpr
}
}
columnExprs.map { c =>
val a = f(c)
Alias(a, a.toString)()
}
}
```
and just remove the old one? There is actually a bug in the old one,
because it doesn't use a resolver to compare column names. On top of that, it
is not uncommon tables have thousands of columns. You have a operation that is
actually O(n^2)...
---
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
