aokolnychyi commented on a change in pull request #3287:
URL: https://github.com/apache/iceberg/pull/3287#discussion_r733226452
##########
File path:
spark/v3.0/spark3/src/main/java/org/apache/iceberg/spark/data/vectorized/ColumnarBatchReader.java
##########
@@ -47,18 +68,57 @@ public final ColumnarBatch read(ColumnarBatch reuse, int
numRowsToRead) {
closeVectors();
}
+ Pair<int[], Integer> rowIdMapping = rowIdMapping(numRowsToRead);
+
for (int i = 0; i < readers.length; i += 1) {
vectorHolders[i] = readers[i].read(vectorHolders[i], numRowsToRead);
int numRowsInVector = vectorHolders[i].numValues();
Preconditions.checkState(
numRowsInVector == numRowsToRead,
"Number of rows in the vector %s didn't match expected %s ",
numRowsInVector,
numRowsToRead);
- arrowColumnVectors[i] =
- IcebergArrowColumnVector.forHolder(vectorHolders[i],
numRowsInVector);
+
+ if (rowIdMapping.second() == 0) {
Review comment:
It is unclear what `rowIdMapping.second() == 0` means.
I think it would be more readable if `rowPosMapping(numRowsToRead)` retuned
null if the delete filter is not set.
Then this could be structured as follows:
```
if (rowPosMapping != null) {
int[] rowPosMap = rowPosMapping.first();
Integer numRows = rowPosMapping.second();
arrowColumnVectors[i] =
ColumnVectorWithFilter.forHolder(vectorHolders[i], rowPosMap, numRows);
} else {
arrowColumnVectors[i] =
IcebergArrowColumnVector.forHolder(vectorHolders[i], numRowsInVector);
}
}
```
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
