[
https://issues.apache.org/jira/browse/PHOENIX-4288?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16267971#comment-16267971
]
ASF GitHub Bot commented on PHOENIX-4288:
-----------------------------------------
Github user JamesRTaylor commented on a diff in the pull request:
https://github.com/apache/phoenix/pull/281#discussion_r153378386
--- Diff:
phoenix-core/src/main/java/org/apache/phoenix/optimize/QueryOptimizer.java ---
@@ -91,8 +91,23 @@ public QueryPlan optimize(PhoenixStatement statement,
SelectStatement select, Co
}
public QueryPlan optimize(QueryPlan dataPlan, PhoenixStatement
statement, List<? extends PDatum> targetColumns, ParallelIteratorFactory
parallelIteratorFactory) throws SQLException {
- List<QueryPlan>plans = getApplicablePlans(dataPlan, statement,
targetColumns, parallelIteratorFactory, true);
- return plans.get(0);
+ List<QueryPlan> plans = getApplicablePlans(dataPlan, statement,
targetColumns, parallelIteratorFactory, false);
+ if (plans.size() == 1) {
+ return plans.get(0);
+ }
+
+ // Get the best plan based on their costs. Costs will be ZERO if
stats are not
+ // available, thus the first plan will be returned.
+ Cost minCost = null;
+ QueryPlan bestPlan = null;
+ for (QueryPlan plan : plans) {
+ Cost cost = plan.getCost();
+ if (minCost == null || cost.compareTo(minCost) < 0) {
+ minCost = cost;
+ bestPlan = plan;
+ }
+ }
+ return bestPlan;
--- End diff --
QueryOptimize.optimize() is not the only entry point to the optimizer. For
example, DeleteCompiler calls both getApplicablePlans() and getBestPlan(). I
think it'd be best if you hid this logic at the end of the private
getApplicablePlans since all calls channel through that. That way all optimizer
decisions will use cost if it's enabled. You could conditional order the plans
based on their cost if cost-based is enabled here:
return hintedPlan != null ? plans : isCostBasedEnabled ?
orderPlansBasedOnCost(plans) : orderPlansBestToWorst(select, plans,
stopAtBestPlan);
> Indexes not used when ordering by primary key
> ---------------------------------------------
>
> Key: PHOENIX-4288
> URL: https://issues.apache.org/jira/browse/PHOENIX-4288
> Project: Phoenix
> Issue Type: Sub-task
> Reporter: Marcin Januszkiewicz
> Assignee: Maryann Xue
> Labels: CostBasedOptimization
>
> We have a table
> CREATE TABLE t (
> rowkey VARCHAR PRIMARY KEY,
> c1 VARCHAR,
> c2 VARCHAR
> )
> which we want to query by doing partial matches on c1, and keep the ordering
> of the source table:
> SELECT rowkey, c1, c2 FROM t where c1 LIKE 'X0%' ORDER BY rowkey;
> We expect most queries to select a small subset of the table, so we create an
> index to speed up searches:
> CREATE LOCAL INDEX t_c1_ix ON t (c1);
> However, this index will not be used since Phoenix will always choose not to
> resort the data.
> In our actual use case, adding index hints is not a practical solution.
> See also discussion at:
> https://lists.apache.org/thread.html/26ab58288eb811d2f074c3f89067163d341e5531fb581f3b2486cf43@%3Cuser.phoenix.apache.org%3E
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
