add row index column only

sezruby · Eunjin Song (from Dev Box) · commit f0c5676b8f88 · 2025-08-14T21:15:24.000Z
diff --git a/spark/src/main/scala/org/apache/spark/sql/delta/PreprocessTableWithDVs.scala b/spark/src/main/scala/org/apache/spark/sql/delta/PreprocessTableWithDVs.scala
@@ -147,7 +147,12 @@ object ScanWithDeletionVectors {
           case o => o
         }
       } else {
-        inputScan.output :+ fileFormat.createFileMetadataCol()
+        val metadataAttr = fileFormat.createFileMetadataCol()
+        val rowIndexOnlyDataType =
+          metadataAttr.dataType.asInstanceOf[StructType].apply(Set(ParquetFileFormat.ROW_INDEX))
+        inputScan.output :+ metadataAttr.copy(
+          dataType = rowIndexOnlyDataType)(
+            exprId = metadataAttr.exprId, qualifier = metadataAttr.qualifier)
       }
     } else {
       inputScan.output

Original file line number	Diff line number	Diff line change
`@@ -147,7 +147,12 @@ object ScanWithDeletionVectors {`
`147`	`147`	`case o => o`
`148`	`148`	`}`
`149`	`149`	`} else {`
`150`		`- inputScan.output :+ fileFormat.createFileMetadataCol()`
	`150`	`+ val metadataAttr = fileFormat.createFileMetadataCol()`
	`151`	`+ val rowIndexOnlyDataType =`
	`152`	`+ metadataAttr.dataType.asInstanceOf[StructType].apply(Set(ParquetFileFormat.ROW_INDEX))`
	`153`	`+ inputScan.output :+ metadataAttr.copy(`
	`154`	`+ dataType = rowIndexOnlyDataType)(`
	`155`	`+ exprId = metadataAttr.exprId, qualifier = metadataAttr.qualifier)`
`151`	`156`	`}`
`152`	`157`	`} else {`
`153`	`158`	`inputScan.output`