(gluten) branch main updated: [VL] test(iceberg): Fix disabled iceberg row group bytes test (#12509)

[yuanzhou]

This is an automated email from the ASF dual-hosted git repository.

zhouyuan pushed a commit to branch main
in repository https://gitbox.apache.org/repos/asf/gluten.git


The following commit(s) were added to refs/heads/main by this push:
     new f734ed79f3 [VL] test(iceberg): Fix disabled iceberg row group bytes 
test (#12509)
f734ed79f3 is described below

commit f734ed79f36b46fe05a5f21e2f010a80fbb1aa59
Author: inf <ialhaz...@gmail.com>
AuthorDate: Mon Jul 20 08:56:37 2026 +0000

    [VL] test(iceberg): Fix disabled iceberg row group bytes test (#12509)
    
    Previously, the row group bytes value was a hard boundary of the 
uncompressed value. Now the implementation causes the buffer to reach or 
slightly exceed the threshold. The tests are updated to guard bigger written 
row group size than configuration.
---
 .../execution/enhanced/VeloxIcebergSuite.scala     | 30 +++++++++++++++++-----
 1 file changed, 23 insertions(+), 7 deletions(-)

diff --git 
a/backends-velox/src-iceberg/test/scala/org/apache/gluten/execution/enhanced/VeloxIcebergSuite.scala
 
b/backends-velox/src-iceberg/test/scala/org/apache/gluten/execution/enhanced/VeloxIcebergSuite.scala
index 837f3917f9..3a8c48d5e9 100644
--- 
a/backends-velox/src-iceberg/test/scala/org/apache/gluten/execution/enhanced/VeloxIcebergSuite.scala
+++ 
b/backends-velox/src-iceberg/test/scala/org/apache/gluten/execution/enhanced/VeloxIcebergSuite.scala
@@ -655,7 +655,7 @@ class VeloxIcebergSuite extends IcebergSuite {
   }
 
   // Ignored due to velox parquet row-group flush semantics change after 
velox#16998.
-  ignore("iceberg parquet writer default row group size test") {
+  test("iceberg parquet writer default row group size test") {
     val table = "iceberg_default_row_group_size"
     val defaultRowGroupBytes = 128L * 1024 * 1024
 
@@ -743,19 +743,35 @@ class VeloxIcebergSuite extends IcebergSuite {
         checkAnswer(
           spark.sql(s"SELECT count(*) FROM $table"),
           Seq(Row(90000L)))
-        val rowGroups = collectRowGroups(table)
+        val rowGroups =
+          collectRowGroups(table).sortBy(info => (info.file, info.ordinal))
+
+        assert(
+          rowGroups.map(_.file).distinct.size == 1,
+          s"Expected one Parquet file, found: 
${rowGroups.map(_.file).distinct}")
 
         assert(
           rowGroups.size == 2,
-          "Expected 2 row groups")
+          s"Expected 2 row groups, found ${rowGroups.size}: $rowGroups")
 
         assert(
-          rowGroups.head.totalByteSize < defaultRowGroupBytes,
-          "Expected row group to contain less than default value")
+          rowGroups.map(_.rowCount).sum == 90000L,
+          s"Expected 90000 rows across all row groups: $rowGroups")
+
+        val firstRowGroup = rowGroups.head
+        val finalRowGroup = rowGroups.last
 
         assert(
-          rowGroups(1).totalByteSize < defaultRowGroupBytes,
-          "Expected row group to contain less than default value")
+          firstRowGroup.compressedSize >= defaultRowGroupBytes,
+          s"Expected the first row group to reach the default row-group size " 
+
+            s"$defaultRowGroupBytes, but found ${firstRowGroup.compressedSize}"
+        )
+
+        assert(
+          finalRowGroup.compressedSize < defaultRowGroupBytes,
+          s"Expected the final row group to be smaller than the default 
row-group " +
+            s"size $defaultRowGroupBytes, but found 
${finalRowGroup.compressedSize}"
+        )
       }
     }
   }


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@gluten.apache.org
For additional commands, e-mail: commits-h...@gluten.apache.org