[
https://issues.apache.org/jira/browse/FLINK-20255?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17751957#comment-17751957
]
Yunhong Zheng commented on FLINK-20255:
---------------------------------------
Hi [~lam167] . By design, your example cannot work. Now, 'UNNEST(xx) AS xx (xx)
ON' can only with condition equals 'TRUE', like 'UNNEST(relatedUserIds) AS t
(relatedUserId) ON TRUE'. Calcite will translate UNNEST as 'Correlate', which
you can treat it as a nested loop join, so you cannot give it a not always true
condition.
For your sql example, you can convert it into a filter condition and add it in
where condition, such as
SELECT *
FROM Messages
CROSS JOIN UNNEST(relatedUserIds) AS t (relatedUserId) where userId =
t.relatedUserId
> Nested decorrelate failed
> -------------------------
>
> Key: FLINK-20255
> URL: https://issues.apache.org/jira/browse/FLINK-20255
> Project: Flink
> Issue Type: Bug
> Components: Table SQL / Planner
> Affects Versions: 1.11.0, 1.12.0
> Reporter: godfrey he
> Priority: Not a Priority
> Labels: auto-deprioritized-major, auto-deprioritized-minor
>
> This issue is from ML
> https://www.mail-archive.com/user@flink.apache.org/msg37746.html
> We can reproduce the issue through the following code
> {code:java}
> @FunctionHint(output = new DataTypeHint("ROW<val STRING>"))
> class SplitStringToRows extends TableFunction[Row] {
> def eval(str: String, separator: String = ";"): Unit = {
> if (str != null) {
> str.split(separator).foreach(s => collect(Row.of(s.trim())))
> }
> }
> }
> object Job {
> def main(args: Array[String]): Unit = {
> val settings =
> EnvironmentSettings.newInstance().useBlinkPlanner().inStreamingMode().build()
> val streamEnv = StreamExecutionEnvironment.getExecutionEnvironment
> val streamTableEnv = StreamTableEnvironment.create(streamEnv, settings)
> streamTableEnv.createTemporarySystemFunction(
> "SplitStringToRows",
> classOf[SplitStringToRows]
> ) // Class defined in previous email
> streamTableEnv.executeSql(
> """
> CREATE TABLE table2 (
> attr1 STRING,
> attr2 STRING,
> attr3 DECIMAL,
> attr4 DATE
> ) WITH (
> 'connector' = 'datagen'
> )""")
> val q2 = streamTableEnv.sqlQuery(
> """
> SELECT
> a.attr1 AS attr1,
> attr2,
> attr3,
> attr4
> FROM table2 p, LATERAL TABLE(SplitStringToRows(p.attr1, ';')) AS
> a(attr1)
> """)
> streamTableEnv.createTemporaryView("view2", q2)
> val q3 =
> """
> SELECT
> w.attr1,
> p.attr3
> FROM table2 w
> LEFT JOIN LATERAL (
> SELECT
> attr1,
> attr3
> FROM (
> SELECT
> attr1,
> attr3,
> ROW_NUMBER() OVER (
> PARTITION BY attr1
> ORDER BY
> attr4 DESC NULLS LAST,
> w.attr2 = attr2 DESC NULLS LAST
> ) AS row_num
> FROM view2)
> WHERE row_num = 1) p
> ON (w.attr1 = p.attr1)
> """
> println(streamTableEnv.explainSql(q3))
> }
> }
> {code}
> The reason is {{RelDecorrelator}} in Calcite can't handle such nested
> decorrelate pattern now
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
