[
https://issues.apache.org/jira/browse/FLINK-16568?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17110993#comment-17110993
]
Jingsong Lee commented on FLINK-16568:
--------------------------------------
It is hard to support this, because the inputs of a aggregation are all
non-constant from calcite in Flink.
> Cannot call Hive UDAF that requires constant arguments
> ------------------------------------------------------
>
> Key: FLINK-16568
> URL: https://issues.apache.org/jira/browse/FLINK-16568
> Project: Flink
> Issue Type: Bug
> Components: Connectors / Hive
> Reporter: Rui Li
> Priority: Major
>
> Calling Hive's {{percentile_approx}} would fail:
> {code}
> hiveShell.execute("create table src (key double,val string)");
> tableEnv.loadModule("hive", new HiveModule());
> tableEnv.sqlQuery("select percentile_approx(key,0.5) from src");
> TableUtils.collectToList(tableEnv.sqlQuery("select
> percentile_approx(key,0.5) from src"));
> {code}
> With error:
> {noformat}
> Caused by: org.apache.hadoop.hive.ql.exec.UDFArgumentTypeException: The
> second argument must be a constant, but decimal(2,1) was passed instead.
> at
> org.apache.hadoop.hive.ql.udf.generic.GenericUDAFPercentileApprox.getEvaluator(GenericUDAFPercentileApprox.java:145)
> at
> org.apache.flink.table.functions.hive.HiveGenericUDAF.createEvaluator(HiveGenericUDAF.java:120)
> at
> org.apache.flink.table.functions.hive.HiveGenericUDAF.init(HiveGenericUDAF.java:92)
> at
> org.apache.flink.table.functions.hive.HiveGenericUDAF.getHiveResultType(HiveGenericUDAF.java:196)
> ... 58 more
> {noformat}
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
