[jira] [Created] (FLINK-21856) Fix the bug of using Python UDF from sub-query as input param of Python UDTF

[Shang Yuanchun (Jira)]

Huang Xingbo created FLINK-21856:
------------------------------------

             Summary: Fix the bug of using Python UDF from sub-query as input param of Python UDTF
                 Key: FLINK-21856
                 URL: https://issues.apache.org/jira/browse/FLINK-21856
             Project: Flink
          Issue Type: Bug
          Components: API / Python
    Affects Versions: 1.12.2, 1.11.3, 1.13.0
            Reporter: Huang Xingbo


This example comes from the user. splitStr is a Python UDTF. train_and_predict is a Python UDF.
{code:python}
t_env.sql_query("""
 select A.hotime ,
         A.before_ta ,
         A.before_rssi ,
         A.after_ta ,
         A.after_rssil ,
         A.nb_tath ,
         A.nb_rssith ,
         nbr_rssi ,
         nbr_ta
 from (SELECT
     hotime ,
     before_ta ,
     before_rssi ,
     after_ta ,
     after_rssil ,
     nb_tath ,
     nb_rssith ,
     train_and_predict(nb_tath, nb_rssith) predict
 FROM
     source) as  A,LATERAL TABLE(splitStr(predict)) as T(nbr_rssi, nbr_ta)
 """)

{code}

The root cause is that `train_and_predict` is a RexCorrelVariable which we don't have relevant logic to deal with.

 A workaround is to use the Table API.





--
This message was sent by Atlassian Jira
(v8.3.4#803005)