[
https://issues.apache.org/jira/browse/FLINK-7145?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16082073#comment-16082073
]
Jark Wu commented on FLINK-7145:
--------------------------------
[~Xiaojun Jin], actually, Flink does support multi-argument UDAGG and
overloaded {{accumulate(...)}} methods, see
{{UserDefinedAggFunctions.WeightedAvg}} in tests which has two
{{accumulate(...)}} methods and each has two input parameters. Flink doesn't
use the {{paramTypes}} to check the operands, but create a custom
{{SqlOperandTypeChecker}} and {{SqlOperandTypeInference}} to check the operands.
Yes, you can post your query & TOPK implementation, so that we can figure out
what goes wrong.
Thanks,
Jark
> Flink SQL API should support multiple parameters for UserDefinedAggFunction
> ---------------------------------------------------------------------------
>
> Key: FLINK-7145
> URL: https://issues.apache.org/jira/browse/FLINK-7145
> Project: Flink
> Issue Type: New Feature
> Components: Table API & SQL
> Reporter: yuemeng
>
> UDAF such as topK and some other udaf with bloom filter need more than one
> parameters ,we should make flink sql support this.
> base on flink sql support DML and multiple parameters udaf,we can execute sql
> like:
> {code}
> CREATE TEMPORARY function 'TOPK' AS
> 'com.xxxx.aggregate.udaf.distinctUdaf.topk.ITopKUDAF';
> INSERT INTO db_sink SELECT id, TOPK(price, 5, 'DESC') FROM kafka_source GROUP
> BY id;
> {code}
--
This message was sent by Atlassian JIRA
(v6.4.14#64029)
