[
https://issues.apache.org/jira/browse/IGNITE-10329?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16710141#comment-16710141
]
Vladimir Ozerov commented on IGNITE-10329:
------------------------------------------
[~pkouznet], my comments:
# {{QueryFactory.selectPersonsByPK}} looks wrong. Range scan by PK ID never
makes sense in real life. We should query only single ID
# {{QueryFactory.selectPersonsJoinOrgWherePersonPK}} - same as p.1
# {{PERSON}} table lacks affinity co-location with {{ORGANIZATION}}. It means
that benchmark will work incorrectly when there are several data nodes. On the
other hand, if we add affinity key, then query by person ID will not be able to
use unicast. So let's keep the schema as is, but throw an exception if there
are several Ignite nodes (is it possible with Yardstick? if not, then how
difficult is it to add?)
# {{QueryFactory.selectAllPersons}} - we should select the whole person, not ID
> Create JDBC "query" and "query join" benchmarks and compare them with
> Postgres and MySQL
> ----------------------------------------------------------------------------------------
>
> Key: IGNITE-10329
> URL: https://issues.apache.org/jira/browse/IGNITE-10329
> Project: Ignite
> Issue Type: Task
> Components: sql, yardstick
> Reporter: Vladimir Ozerov
> Assignee: Pavel Kuznetsov
> Priority: Major
> Fix For: 2.8
>
>
> Currently we have {{IgniteSqlQueryBenchmark}} and
> {{IgniteSqlQueryJoinBenchmark}} benchmarks which query data over salary range
> and optionally joins it with second table. Let's create a set of similar
> benchmarks which will use JDBC to load and query data, and execute them
> against one-node Ignite cluster, MySQL and Postgres.
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)
