[ https://issues.apache.org/jira/browse/CALCITE-5163 ]
Xurenhe deleted comment on CALCITE-5163:
----------------------------------
was (Author: wojustme):
Hi guys, PR has been ready, welcome to review it.
https://github.com/apache/calcite/pull/2812
> MysqlSqlDialect unparse LISTAGG aggregate function error
> --------------------------------------------------------
>
> Key: CALCITE-5163
> URL: https://issues.apache.org/jira/browse/CALCITE-5163
> Project: Calcite
> Issue Type: Bug
> Components: core
> Reporter: Xurenhe
> Assignee: Xurenhe
> Priority: Major
>
> CALCITE-4349 supports GROUP_CONCAT aggregate function. [The
> code|https://github.com/apache/calcite/blob/7c1e2746b94a50f0bb08571287acb5327fdc1b16/core/src/main/java/org/apache/calcite/sql2rel/SqlToRelConverter.java#L5746]
> translated *GROUP_CONCAT* to *LISTAGG* during sql_to_rel.
> But, *MysqlSqlDialect* keep *LISTAGG* after uparsing query's relnode, the
> result sql couldn't be executed in MySQL engine.
>
> I expect the following test case to be successful, but it's failed.
> {code:java}
> // test in org.apache.calcite.rel.rel2sql.RelToSqlConverterTest
> @Test void testMySqlGroupConcat() {
> final String query = "select\n"
> + "listagg(distinct \"product_name\", ',') within group(order by
> \"cases_per_pallet\")"
> + "from \"product\"\n"
> + "group by \"product_id\"\n";
> final String expected = "SELECT GROUP_CONCAT(DISTINCT `product_name` ORDER
> BY "
> + "`cases_per_pallet` IS NULL, `cases_per_pallet` SEPARATOR ',')\n"
> + "FROM `foodmart`.`product`\n"
> + "GROUP BY `product_id`";
> sql(query).withMysql().ok(expected);
> }{code}
--
This message was sent by Atlassian Jira
(v8.20.7#820007)
