[
https://issues.apache.org/jira/browse/CALCITE-7050?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17970190#comment-17970190
]
suibianwanwan commented on CALCITE-7050:
----------------------------------------
Regarding rewriting plans in RelToSqlConverter: as far as I know, we already do
that:
{code:java}
if (!this.dialect.supportsGroupByLiteral()) {
HepProgramBuilder hepProgramBuilder = new HepProgramBuilder();
hepProgramBuilder.addRuleInstance(
AggregateProjectConstantToDummyJoinRule.Config.DEFAULT.toRule());
HepPlanner hepPlanner = new HepPlanner(hepProgramBuilder.build());
hepPlanner.setRoot(r);
best = hepPlanner.findBestExp();
} else {
best = r;
}
{code}
> Invalid unparse for FULL JOIN in MySQLDialect
> ----------------------------------------------
>
> Key: CALCITE-7050
> URL: https://issues.apache.org/jira/browse/CALCITE-7050
> Project: Calcite
> Issue Type: Bug
> Reporter: xiong duan
> Assignee: Yu Xu
> Priority: Major
> Labels: pull-request-available
>
> The SQL in MySQL:
> {code:java}
> SELECT * FROM test002 as table1 full join test002 as table2 on table1.id1 =
> table2.id1; {code}
> throws exception:
> {code:java}
> SQL ERROR [1064] [42000]: You have an error in your SQL syntax; check the
> manual that corresponds to your MySQL server version for the right syntax to
> use near 'full join test002 as table2 on table1.id1 = table2.id1 {code}
> Test case in RelToSqlConverterTest:
> {code:java}
> @Test void testFullJoin() {
> String sql = "select *\n"
> + "from \"store\" as s\n"
> + "full join \"employee\" as e on true\n";
> final String expectedMysql = "SELECT *\n"
> + "FROM `foodmart`.`store`\n"
> + "FULL JOIN `foodmart`.`employee` ON TRUE";
> sql(sql).withMysql().ok(expectedMysql);
> } {code}
> We already have `supportsJoinType` when the jointype is JoinRelTpte.FULL
> return false.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)