[
https://issues.apache.org/jira/browse/PHOENIX-4611?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16368026#comment-16368026
]
Maryann Xue commented on PHOENIX-4611:
--------------------------------------
Sorry, [~jamestaylor], I made a mistake in my last comment.
It was caused by the group-by expression re-ordering in GroupByCompiler as
follows:
{code:java}
Collections.sort(groupBys, new Comparator<Pair<Integer,Expression>>() {
@Override
public int compare(Pair<Integer,Expression> gb1, Pair<Integer,Expression>
gb2) {
Expression e1 = gb1.getSecond();
Expression e2 = gb2.getSecond();
PDataType t1 = e1.getDataType();
PDataType t2 = e2.getDataType();
boolean isFixed1 = t1.isFixedWidth();
boolean isFixed2 = t2.isFixedWidth();
boolean isFixedNullable1 = e1.isNullable() &&isFixed1;
boolean isFixedNullable2 = e2.isNullable() && isFixed2;
boolean oae1 = onlyAtEndType(e1);
boolean oae2 = onlyAtEndType(e2);
if (oae1 == oae2) {
if (isFixedNullable1 == isFixedNullable2) {
if (isFixed1 == isFixed2) {
// Not strictly necessary, but forces the order to match
the schema
// column order (with PK columns before value columns).
//return o1.getColumnPosition() - o2.getColumnPosition();
return gb1.getFirst() - gb2.getFirst();
} else if (isFixed1) {
return -1;
} else {
return 1;
}
} else if (isFixedNullable1) {
return 1;
} else {
return -1;
}
} else if (oae1) {
return 1;
} else {
return -1;
}
}
});{code}
And it did not affect the ResultSet. It just changed the execution plan a
little bit, with two expressions swapped. Can you verify that it was the right
behavior after PHOENIX-2566? If yes, we can revert
https://git-wip-us.apache.org/repos/asf?p=phoenix.git;a=blobdiff;f=phoenix-core/src/main/java/org/apache/phoenix/compile/TupleProjectionCompiler.java;h=fccded2a896855a2a01d727b992f954a1d3fa8ab;hp=d0b900c1a9c21609b89065307433a0d37b12b72a;hb=82ba1417fdd69a0ac57cbcf2f2327d4aa371bcd9;hpb=e126dd1dda5aa80e8296d3b0c84736b22b658999
and change tests accordingly.
> Not nullable column impact on join query plans
> ----------------------------------------------
>
> Key: PHOENIX-4611
> URL: https://issues.apache.org/jira/browse/PHOENIX-4611
> Project: Phoenix
> Issue Type: Bug
> Reporter: James Taylor
> Priority: Major
>
> With PHOENIX-2566, there's a subtle change in projected tables in that a
> column may end up being not nullable where as before it was nullable when the
> family name is not null. I've kept the old behavior with
> [this|https://git-wip-us.apache.org/repos/asf?p=phoenix.git;a=blobdiff;f=phoenix-core/src/main/java/org/apache/phoenix/compile/TupleProjectionCompiler.java;h=fccded2a896855a2a01d727b992f954a1d3fa8ab;hp=d0b900c1a9c21609b89065307433a0d37b12b72a;hb=82ba1417fdd69a0ac57cbcf2f2327d4aa371bcd9;hpb=e126dd1dda5aa80e8296d3b0c84736b22b658999]
> commit, but would you mind confirming what the right thing to do is,
> [~maryannxue]?
> Without this change, the explain plan changes in
> SortMergeJoinMoreIT.testBug2894() and the assert fails. Looks like the
> compiler ends up changing the row ordering.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
