Tim Armstrong has posted comments on this change. Change subject: Impala-3286: Prefetching For Phj Probing. ......................................................................
Patch Set 3: (2 comments) http://gerrit.cloudera.org:8080/#/c/2959/3/be/src/exec/partitioned-hash-join-node-ir.cc File be/src/exec/partitioned-hash-join-node-ir.cc: Line 417: int PartitionedHashJoinNode::ProcessProbeBatch( > Tried moving it out of this file but had trouble linking impalad. It may be Oh right, yeah, we probably just need to explicitly instantiate the templates at the bottom of the file. http://gerrit.cloudera.org:8080/#/c/2959/3/be/src/runtime/row-batch.h File be/src/runtime/row-batch.h: Line 151: int limit = FIXED_LEN_BUFFER_LIMIT) : > That's the theoretical maximum number of rows if we only have a column of T It's not a hard limit - it's possible to create a RowBatch with arbitrary capacity (e.g. BTS::GetRows()). The batch is only shrunk down if the caller calls ResizeAndAllocateTupleBuffer(). We can also have row batches with no materialized columns. I think the only hard upper bound is INT_MAX. Maybe rename FIXED_LEN_BUFFER_LIMIT to FIXED_LEN_BUFFER_SOFT_LIMIT to avoid confusion? I originally named the variable. -- To view, visit http://gerrit.cloudera.org:8080/2959 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: Ib42b93d99d09c833571e39d20d58c11ef73f3cc0 Gerrit-PatchSet: 3 Gerrit-Project: Impala Gerrit-Branch: cdh5-trunk Gerrit-Owner: Michael Ho <[email protected]> Gerrit-Reviewer: Dan Hecht <[email protected]> Gerrit-Reviewer: Michael Ho <[email protected]> Gerrit-Reviewer: Mostafa Mokhtar <[email protected]> Gerrit-Reviewer: Tim Armstrong <[email protected]> Gerrit-HasComments: Yes
