[
https://issues.apache.org/jira/browse/FLINK-4294?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15580335#comment-15580335
]
ASF GitHub Bot commented on FLINK-4294:
---------------------------------------
Github user twalthr commented on a diff in the pull request:
https://github.com/apache/flink/pull/2319#discussion_r83556835
--- Diff: docs/dev/table_api.md ---
@@ -1656,6 +1656,29 @@ temporalOverlaps(TIMEPOINT, TEMPORAL, TIMEPOINT,
TEMPORAL)
</td>
</tr>
+ <tr>
+ <td>
+ {% highlight java %}
+ANY.flatten()
+{% endhighlight %}
+ </td>
+ <td>
+ <p>Converts a Flink composite type (such as Tuple, POJO, etc.) and
all of its subtypes into a flat representation where every subtype is a
separate field.</p>
+ </td>
+ </tr>
+
+ <tr>
+ <td>
+ {% highlight java %}
+COMPOSITE.get(STRING)
--- End diff --
I thought we only agreed that `getField` is too long. Personally I like
this approach more because it can be found in the expression DSL and has a
Scala doc; and it doesn't add further case distinction logic to the validation
layer. Since "$" is valid part of Java identifier we would also restrict the
field naming of POJOs. If there is a consensus I can change it to
"field$substring".
> Allow access of composite type fields
> -------------------------------------
>
> Key: FLINK-4294
> URL: https://issues.apache.org/jira/browse/FLINK-4294
> Project: Flink
> Issue Type: New Feature
> Components: Table API & SQL
> Reporter: Timo Walther
> Assignee: Timo Walther
>
> Currently all Flink CompositeTypes are treated as GenericRelDataTypes. It
> would be better to access individual fields of composite types, too. e.g.
> {code}
> SELECT composite.name FROM composites
> SELECT tuple.f0 FROM tuples
> 'f0.getField(0)
> {code}
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)