[
https://issues.apache.org/jira/browse/TAJO-1658?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14604566#comment-14604566
]
ASF GitHub Bot commented on TAJO-1658:
--------------------------------------
Github user hyunsik commented on a diff in the pull request:
https://github.com/apache/tajo/pull/613#discussion_r33419852
--- Diff:
tajo-storage/tajo-storage-common/src/main/java/org/apache/tajo/storage/MergeScanner.java
---
@@ -165,7 +167,7 @@ public boolean isSelectable() {
}
@Override
- public void setSearchCondition(Object expr) {
+ public void setFilter(EvalNode filter) {
--- End diff --
I'll fix it for current implementation. However, we need to know its
ambiguity, and the solution should be depend on the final purpose of
MergeScanner.
According to its implementation, MergeScanner is chosen if the number of
``fragments`` is more than one.
If MergeScanner is only for intermediate shuffled data, this method should
throw an exception. Otherwise, we should implement to push down a filter to its
scannner elements.
> Filter push down to underlying storages
> ---------------------------------------
>
> Key: TAJO-1658
> URL: https://issues.apache.org/jira/browse/TAJO-1658
> Project: Tajo
> Issue Type: Sub-task
> Components: physical operator, storage
> Reporter: Hyunsik Choi
> Assignee: Hyunsik Choi
> Fix For: 0.11.0
>
>
> Some storages can filter unqualified rows. For it, SeqScanner should support
> the filter push down to underlying scanners.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
