alamb commented on code in PR #11966:
URL: https://github.com/apache/datafusion/pull/11966#discussion_r1716727402
##########
datafusion/physical-plan/src/coalesce_batches.rs:
##########
@@ -364,90 +393,60 @@ impl BatchCoalescer {
Arc::clone(&self.schema)
}
- /// Add a batch, returning a batch if the target batch size or limit is
reached
- fn push_batch(&mut self, batch: RecordBatch) ->
Result<Option<RecordBatch>> {
- // discard empty batches
- if batch.num_rows() == 0 {
- return Ok(None);
- }
+ fn push_batch(&mut self, batch: RecordBatch) -> CoalescerState {
+ let batch = gc_string_view_batch(&batch);
Review Comment:
gc'ing basically compacts string data together so it requires less memory to
store but requires a copy of the values
I think the implication of doing this gc before checking for limit is that
we may copy the last batch worth of data prior to a limit being hit just to
throw it away. However, the plan is about to shut down anyways (will switch to
`CoalesceBatchesStreamState::Exhausted`) so I don't think this is a major
problem
--
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: github-unsubscr...@datafusion.apache.org
For queries about this service, please contact Infrastructure at:
us...@infra.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: github-unsubscr...@datafusion.apache.org
For additional commands, e-mail: github-h...@datafusion.apache.org
