问题描述
我想创建一个包含MATCH_RECOGNIZE语句的视图。这会遇到以下错误：
org.apache.flink.table.api.SqlParserException: SQL parse failed. Encountered "MATCH_RECOGNIZE" at line 3,column 79.
Was expecting one of:
<EOF>
"EXCEPT" ...
"FETCH" ...
"GROUP" ...
"HAVING" ...
"INTERSECT" ...
"LIMIT" ...
"OFFSET" ...
"ORDER" ...
"MINUS" ...
"TABLESAMPLE" ...
"UNION" ...
"WHERE" ...
"WINDOW" ...
"(" ...
"NATURAL" ...
"JOIN" ...
"INNER" ...
"LEFT" ...
"RIGHT" ...
"FULL" ...
"CROSS" ...
"," ...
"OUTER" ...
如果您想重现此问题，可以例如请遵循Ververica training，该内容也涵盖了MATCH_RECOGNIZE。
初始设置后，创建以下视图：
CREATE VIEW `RideView`
AS
SELECT rideId,TIMESTAMPDIFF(MINUTE,startT,endT) AS durationMin
FROM Rides
MATCH_RECOGNIZE (
PARTITION BY rideId
ORDER BY rideTime
MEASURES
S.rideTime AS startT,E.rideTime AS endT
AFTER MATCH SKIP PAST LAST ROW
PATTERN (S E)
DEFINE
S AS S.isStart,E AS NOT E.isStart
);
仅运行SELECT * FROM RideView就会触发问题。
这并不意外，因为Flink documentation表示
当前，MATCH_RECOGNIZE子句只能应用于追加表。此外，它也总是产生一个追加表。 （https://ci.apache.org/projects/flink/flink-docs-stable/dev/table/streaming/match_recognize.html#sql-semantics）。
是否有任何方法可以克服此限制，并与CREATE VIEW AS <query>结合使用MATCH_RECOGNIZE？
解决方法
暂无找到可以解决该程序问题的有效方法，小编努力寻找整理中！
如果你已经找到好的解决方法，欢迎将解决方案带上本链接一起发送给小编。
小编邮箱:dio#foxmail.com (将#修改为@）