[
https://issues.apache.org/jira/browse/FLINK-6124?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15935658#comment-15935658
]
ASF GitHub Bot commented on FLINK-6124:
---------------------------------------
Github user shaoxuan-wang commented on the issue:
https://github.com/apache/flink/pull/3579
@twalthr, I saw you have removed the CHAR type when you merge this. Just
want to clarify. In FLINK-3916, in method typeInfoToSqlTypeName, you have added
```
case CHAR_TYPE_INFO | CHAR_VALUE_TYPE_INFO =>
throw TableException("Character type is not supported.")
```
TableAPI is open for user to pass CHAR type data,
```
val data = List(
(1L, 1, 1d, 1f, new BigDecimal("1"), 'a', "Hi")
...
.toTable(tEnv, 'long, 'int, 'double, 'float, 'bigdec, 'string)
...
select('char.count)
```
I am curious the reason why we do not support CHAR type explicitly.
Also, as I suggested early, we'd better to have all build-in aggregates to
support the retraction method. We should add string type for
Max/MinWithRetractAggFunction. What do you think?
> support max/min aggregations for string type
> --------------------------------------------
>
> Key: FLINK-6124
> URL: https://issues.apache.org/jira/browse/FLINK-6124
> Project: Flink
> Issue Type: Improvement
> Components: Table API & SQL
> Reporter: Zhenghua Gao
> Assignee: Zhenghua Gao
> Fix For: 1.3.0
>
>
> Recently when I port some query to Flink SQL, I found currently min/max
> aggregations on string type is not supported and should be added.
> When min/max aggregations are used on string column, return min/max value by
> lexicographically order.
>
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
