[
https://issues.apache.org/jira/browse/DRILL-4823?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15633295#comment-15633295
]
ASF GitHub Bot commented on DRILL-4823:
---------------------------------------
Github user jinfengni commented on a diff in the pull request:
https://github.com/apache/drill/pull/560#discussion_r86382118
--- Diff:
exec/java-exec/src/main/java/org/apache/drill/exec/ops/FragmentContext.java ---
@@ -445,11 +447,16 @@ public PartitionExplorer getPartitionExplorer() {
}
@Override
- public ValueHolder getConstantValueHolder(String value,
Function<DrillBuf, ValueHolder> holderInitializer) {
- ValueHolder valueHolder = constantValueHolderCache.get(value);
+ public ValueHolder getConstantValueHolder(String value, MinorType type,
Function<DrillBuf, ValueHolder> holderInitializer) {
--- End diff --
it would be nice if we could share the common codes of
getConstantValueHolder() in QueryContext/FragmentContext, and wrap them in a
helper function. But this is optional.
> Fix OOM while trying to prune partitions with reasonable data size
> ------------------------------------------------------------------
>
> Key: DRILL-4823
> URL: https://issues.apache.org/jira/browse/DRILL-4823
> Project: Apache Drill
> Issue Type: Bug
> Components: Functions - Drill, Query Planning & Optimization
> Affects Versions: 1.6.0
> Reporter: Arina Ielchiieva
> Assignee: Jinfeng Ni
> Fix For: 1.9.0
>
>
> _Example query:_
> {code:sql}
> select '/'||dir0||'/'||dir1||'/'||dir2||'/'||dir3 , count(*)
> FROM dfs.`/path/to/parquet/files`
> WHERE ('/'||dir0||'/'||dir1||'/'||dir2||'/'||dir3) IN (
> '/2015/11/30',
> '//2015/09/01',
> '/2015/09/02',
> '/2015/09/03',
> '/2015/09/04',
> '/2015/09/09',
> '/2016/03/30'
> )
> group by '/'||dir0||'/'||dir1||'/'||dir2||'/'||dir3
> order by 1;
> {code}
> _Error:_
> OOM while trying to prune partitions:
> {noformat}
> org.apache.drill.exec.exception.OutOfMemoryException: Unable to allocate
> buffer of size 256 due to memory limit. Current allocation: 5242880
> at
> org.apache.drill.exec.memory.BaseAllocator.buffer(BaseAllocator.java:216)
> ~[drill-memory-base-1.6.0.jar:1.6.0]
> at
> org.apache.drill.exec.ops.BufferManagerImpl.getManagedBuffer(BufferManagerImpl.java:60)
> ~[drill-java-exec-1.6.0.jar:1.6.0]
> at
> org.apache.drill.exec.ops.BufferManagerImpl.getManagedBuffer(BufferManagerImpl.java:56)
> ~[drill-java-exec-1.6.0.jar:1.6.0]
> at
> org.apache.drill.exec.ops.QueryContext.getManagedBuffer(QueryContext.java:241)
> ~[drill-java-exec-1.6.0.jar:1.6.0]
> at
> org.apache.drill.exec.expr.fn.interpreter.InterpreterEvaluator$EvalVisitor.getManagedBufferIfAvailable(InterpreterEvaluator.java:158)
> ~[drill-java-exec-1.6.0.jar:1.6.0]
> {noformat}
> _Cause:_
> Interpreter always asks for a new buffer to hold varchar/varbinary or decimal
> constant values. That's why the memory size required would be proportion to #
> of constant expressions multiplied by # of input rows (partition). This is
> different from evaluation from run-time generated where constant expression
> will be evaluated once and use only one buffer per value.
> _Fix:_
> To use one buffer for each unique constant value in query.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
