[jira] [Created] (SPARK-21513) SQL to_json should support all column types

["Aaron Davidson (JIRA)" <ji...@apache.org>]

Aaron Davidson created SPARK-21513:
--------------------------------------

             Summary: SQL to_json should support all column types
                 Key: SPARK-21513
                 URL: https://issues.apache.org/jira/browse/SPARK-21513
             Project: Spark
          Issue Type: Improvement
          Components: SQL
    Affects Versions: 2.2.0
            Reporter: Aaron Davidson


The built-in SQL UDF "to_json" currently supports serializing StructType columns, as well as Arrays of StructType columns. If you attempt to use it on a different type, for example a map, you get an error like this:

{code}
AnalysisException: cannot resolve 'structstojson(`tags`)' due to data type mismatch: Input type map<string,string> must be a struct or array of structs.;;
{code}

This limitation seems arbitrary; if I were to go through the effort of enclosing my map in a struct, it would be serializable. Same thing with any other non-struct type.

Therefore the desired improvement is to allow to_json to operate directly on any column type. The associated code is [here](https://github.com/apache/spark/blob/86174ea89b39a300caaba6baffac70f3dc702788/sql/catalyst/src/main/scala/org/apache/spark/sql/catalyst/expressions/jsonExpressions.scala#L653).



--
This message was sent by Atlassian JIRA
(v6.4.14#64029)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org