Alex Behm has posted comments on this change.

Change subject: PREVIEW IMPALA-2521: Add clustered hint to insert statements

Patch Set 1:

File fe/src/main/java/org/apache/impala/planner/

Line 269:    * clustered/noclustered plan hint. The clustering re-order the 
data in 'inputFragment'
Better to be specific: Say that it adds a sort node that orders on the 
partition columns of the target table.

Line 272:   // TODO: this function shares some code with 
QueryStmt::createSortTupleInfo(). However
You could move most of the code into SortInfo.createSortTupleInfo(List<Expr> 

Line 309:     sortTupleDesc.setIsMaterialized(true);
To indicate that this tuple is a physical tuple needed during runtime execution 
and not a "virtual" tuple like e.g. the tuple of an inline view (which never 
gets materialized).

Line 324:     // TODO Why is this needed here but not in If it 
is omitted, the query
Before plan generation we call QueryStmt.materializeRequiredSlots() to make 
sure all slots needed to evaluate that QueryStmt are marked as materialized. 
However, we have no such mechanism here because we are adding a plan node that 
does not come from a SQL clause, so we must mark the slots as materialized 
File fe/src/main/java/org/apache/impala/planner/

Line 144:         rootFragment = distributedPlanner.addClusteringFragment(
1. we should also do this for single-node plans, so adding the clustering 
doesn't really belong in the DistributedPlanner
2. we are not adding a new fragment, so I'd rephrase this addClusteringNode() 
or similar

To view, visit
To unsubscribe, visit

Gerrit-MessageType: comment
Gerrit-Change-Id: I412153bd8435d792bd61dea268d7a3b884048f14
Gerrit-PatchSet: 1
Gerrit-Project: Impala-ASF
Gerrit-Branch: master
Gerrit-Owner: Lars Volker <>
Gerrit-Reviewer: Alex Behm <>
Gerrit-Reviewer: Lars Volker <>
Gerrit-Reviewer: Marcel Kornacker <>
Gerrit-HasComments: Yes

Reply via email to