viirya commented on code in PR #11218:
URL: https://github.com/apache/datafusion/pull/11218#discussion_r1682243474
##########
datafusion/physical-plan/src/joins/sort_merge_join.rs:
##########
@@ -1502,13 +1572,45 @@ fn get_buffered_columns(
buffered_data: &BufferedData,
buffered_batch_idx: usize,
buffered_indices: &UInt64Array,
-) -> Result<Vec<ArrayRef>, ArrowError> {
- buffered_data.batches[buffered_batch_idx]
- .batch
- .columns()
- .iter()
- .map(|column| take(column, &buffered_indices, None))
- .collect::<Result<Vec<_>, ArrowError>>()
+) -> Result<Vec<ArrayRef>> {
+ get_buffered_columns_from_batch(
+ &buffered_data.batches[buffered_batch_idx],
+ buffered_indices,
+ )
+}
+
+#[inline(always)]
+fn get_buffered_columns_from_batch(
+ buffered_batch: &BufferedBatch,
+ buffered_indices: &UInt64Array,
+) -> Result<Vec<ArrayRef>> {
+ match (&buffered_batch.spill_file, &buffered_batch.batch) {
+ // In memory batch
+ (None, Some(batch)) => Ok(batch
+ .columns()
+ .iter()
+ .map(|column| take(column, &buffered_indices, None))
+ .collect::<Result<Vec<_>, ArrowError>>()
+ .map_err(Into::<DataFusionError>::into)?),
+ // If the batch was spilled to disk, less likely
+ (Some(spill_file), None) => {
+ let mut buffered_cols: Vec<ArrayRef> =
+ Vec::with_capacity(buffered_indices.len());
+
+ let file = BufReader::new(File::open(spill_file.path())?);
+ let reader = FileReader::try_new(file, None)?;
+
+ for batch in reader {
+ batch?.columns().iter().for_each(|column| {
+ buffered_cols.extend(take(column, &buffered_indices, None))
+ });
+ }
+
+ Ok(buffered_cols)
+ }
+ // Invalid combination
+ _ => internal_err!("Buffered batch spill status is in the inconsistent
state."),
Review Comment:
Can we add more details in the error message to help debugging? For example,
what the inconsistent state is, i.e., spilling file exists but in-memory batch
also exists, both are not existing.
--
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]
