Github user kiszk commented on a diff in the pull request:
https://github.com/apache/spark/pull/11984#discussion_r57928597
--- Diff:
sql/core/src/main/scala/org/apache/spark/sql/execution/columnar/GenerateColumnAccessor.scala
---
@@ -114,6 +123,57 @@ object GenerateColumnAccessor extends
CodeGenerator[Seq[DataType], ColumnarItera
(createCode, extract + patch)
}.unzip
+ val accessorCode = accessorClasses.map { case (accessorCls,
shortAccCls) =>
+ s"""
+ private $accessorCls get${shortAccCls}(int idx) {
+ byte[] buffer = batch.buffers()[columnIndexes[idx]];
+ return new
$accessorCls(ByteBuffer.wrap(buffer).order(nativeOrder));
+ }
+ """
+ }
+ val accessorStructCode = accessorStructClasses.map {
+ case ((accessorCls, dtCls), (shortAccCls, shortDTCls)) =>
+ s"""
+ private $accessorCls get${shortAccCls}_${shortDTCls}(int idx) {
+ byte[] buffer = batch.buffers()[columnIndexes[idx]];
+ return new
$accessorCls(ByteBuffer.wrap(buffer).order(nativeOrder),
+ (${dtCls}) columnTypes[idx]);
+ }
+ """
+ }
+
+ /* 4000 = 64000 bytes / 16 (up to 16 bytes per one call)) */
+ val numberOfStatementsThreshold = 4000
--- End diff --
I confirmed this by running the following program:
```
val df = sc.parallelize(1 to 100).toDF()
val aggr = {1 to 3000}.map(colnum =>
avg(df.col("_1")).as(s"col_$colnum"))
val res = df.groupBy("_1").agg(count("_1"), aggr: _*).cache()
var i = 0
for (i <- 0 to 110)
res.collect()
```
---
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
