[
https://issues.apache.org/jira/browse/SPARK-16892?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15970348#comment-15970348
]
Hyukjin Kwon commented on SPARK-16892:
--------------------------------------
Maybe you are looking for somethine like this?
{code}
scala> Seq(Tuple1(Array(Array(1, 2), Array(3, 4)))).toDS.map(r =>
r._1.flatten).show()
+------------+
| value|
+------------+
|[1, 2, 3, 4]|
+------------+
{code}
> flatten function to get flat array (or map) column from array of array (or
> array of map) column
> -----------------------------------------------------------------------------------------------
>
> Key: SPARK-16892
> URL: https://issues.apache.org/jira/browse/SPARK-16892
> Project: Spark
> Issue Type: New Feature
> Components: SQL
> Reporter: Kapil Singh
>
> flatten(input)
> Converts input of array of array type into flat array type by inserting
> elements of all element arrays into single array. Example:
> input: [[1, 2, 3], [4, 5], [-1, -2, 0]]
> output: [1, 2, 3, 4, 5, -1, -2, 0]
> Converts input of array of map type into flat map type by inserting key-value
> pairs of all element maps into single map. Example:
> input: [(1 -> "one", 2 -> "two"), (0 -> "zero"), (4 -> "four")]
> output: (1 -> "one", 2 -> "two", 0 -> "zero", 4 -> "four")
--
This message was sent by Atlassian JIRA
(v6.3.15#6346)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
