[
https://issues.apache.org/jira/browse/IMPALA-7754?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16680342#comment-16680342
]
Paul Rogers commented on IMPALA-7754:
-------------------------------------
Another misfire, in {{PlannerTest}}, {{constant-propagation.test}}:
{noformat}
# Non-trivial expr substitution (non-constant)
select count(*) from
(select * from functional.alltypes where int_col = 10) T
where T.int_col = 10 and T.int_col > 1 and T.int_col = 10 and T.int_col =
T.int_col and
(coalesce(NULL, T.int_col) + random() * T.tinyint_col = 100 OR
coalesce(NULL, T.int_col) + T.int_col = 20)
---- PLAN
PLAN-ROOT SINK
|
01:AGGREGATE [FINALIZE]
| output: count(*)
|
00:SCAN HDFS [functional.alltypes]
partitions=24/24 files=24 size=478.45KB
predicates: int_col = 10, TRUE OR 10 + random() *
functional.alltypes.tinyint_col = 100
{noformat}
Should be:
{noformat}
00:SCAN HDFS [functional.alltypes]
partitions=24/24 files=24 size=478.45KB
predicates: int_col = 10
{noformat}
> Expressions sometimes not re-analyzed after rewrite
> ---------------------------------------------------
>
> Key: IMPALA-7754
> URL: https://issues.apache.org/jira/browse/IMPALA-7754
> Project: IMPALA
> Issue Type: Bug
> Components: Frontend
> Affects Versions: Impala 3.0
> Reporter: Paul Rogers
> Priority: Major
>
> The analyzer has a chain of rules which fire in order without (as noted
> above) repeats. The result of rule A (rewriting conditional functions) is fed
> into rule B (simplify CASE). Each rule requires that analysis be done so that
> attributes of expressions can be picked out.
> As it turns out, in the current code, this is rather ad-hoc. The
> {{SimplifyConditionalsRule}} re-analyzes its result as part of the fix for
> IMPALA-5125, but others do not, leading to optimizations not working. In
> particular, in a chain of rewrites for {{IS DISTINCT FROM}}, certain rules
> didn't fire because previous rules left new expressions in an un-analyzed
> state. This is a bug.
> The fix is to analyze the result any time a rule fires, before passing the
> result to the next rule.
> {code:java}
> private Expr applyRuleBottomUp(Expr expr, ExprRewriteRule rule, Analyzer
> analyzer)
> throws AnalysisException {
> ...
> Expr rewrittenExpr = rule.apply(expr, analyzer);
> if (rewrittenExpr != expr) {
> ++numChanges_;
> rewrittenExpr.analyze(analyzer); // Add me!
> }}
> return rewrittenExpr;
> }
> {code}
> There are several places that the above logic appears: make the change in all
> of them.
> Then, in rules that simply refused to run if an expression is to analyzed:
> {code:java}
> public class SimplifyDistinctFromRule implements ExprRewriteRule {
> public Expr apply(Expr expr, Analyzer analyzer) {
> if (!expr.isAnalyzed()) return expr;
> {code}
> Replace this with an assertion that analysis must have been done:
> {code:java}
> public class SimplifyDistinctFromRule implements ExprRewriteRule {
> public Expr apply(Expr expr, Analyzer analyzer) {
> assert expr.isAnalyzed();
> {code}
> To be safe, the assertion fires only in "debug" mode, not in production.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
