[
https://issues.apache.org/jira/browse/TAJO-1548?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14500208#comment-14500208
]
ASF GitHub Bot commented on TAJO-1548:
--------------------------------------
Github user charsyam commented on the pull request:
https://github.com/apache/tajo/pull/531#issuecomment-94029362
Hi, @blrunner
CHAR type can have length. :)
But if we make CHAR Column using this command
in
tajo-storage/tajo-storage-hdfs/src/test/java/org/apache/tajo/storage/parquet/TestReadWrite.java
```java
columns.add(new Column("mychar", Type.CHAR));
```
but it's length is 0, because it is just using CatalogUtil.newSimpleDataType
and Scheme class already use condition code in addColumn method.
But I think it can cause some mistakes easily.
So I think it is better to move this condition code into
CatalogUtil.newSimpleDataType
What do you think?
> Refactoring condition code for CHAR into CatalogUtil
> ----------------------------------------------------
>
> Key: TAJO-1548
> URL: https://issues.apache.org/jira/browse/TAJO-1548
> Project: Tajo
> Issue Type: Improvement
> Reporter: DaeMyung Kang
> Assignee: DaeMyung Kang
> Priority: Minor
>
> Currently Schema's addColumn is this
> ```java
> public synchronized Schema addColumn(String name, Type type) {
> if (type == Type.CHAR) {
> return addColumn(name, CatalogUtil.newDataTypeWithLen(type, 1));
> }
> return addColumn(name, CatalogUtil.newSimpleDataType(type));
> }
> ```
> I think it is better condition code into CatalogUtil.newSimpleDataType.
> because some testcode make columns just using CatalogUtil.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
