[jira] [Commented] (SPARK-22241) Apache spark giving InvalidSchemaException: Cannot write a schema with an empty group: optional group element {

["Ritika Maheshwari (JIRA)" <ji...@apache.org>]

    [ https://issues.apache.org/jira/browse/SPARK-22241?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16199724#comment-16199724 ] 

Ritika Maheshwari commented on SPARK-22241:
-------------------------------------------

I know parquet does not allow empty struct types. But is this something that the Encoder should handle when generating the schema

> Apache spark giving InvalidSchemaException: Cannot write a schema with an empty group: optional group element {
> ---------------------------------------------------------------------------------------------------------------
>
>                 Key: SPARK-22241
>                 URL: https://issues.apache.org/jira/browse/SPARK-22241
>             Project: Spark
>          Issue Type: Bug
>          Components: Spark Core
>    Affects Versions: 2.2.0
>            Reporter: Ritika Maheshwari
>            Priority: Minor
>
> I have a bean which has field of type Arraylist of Doubles. Then I do the following
> Encoder<T> beanEncoder = Encoders.bean(jClass);
> Dataset<T> df = spark.createDataset( Collections.singletonList((T)extractedObj),beanEncoder);
> The schema generated says:
> |-- pixelSpacing: array (nullable = true)
> |-- element: struct (containsNull = true)
> Now I try to save this Dataset as parquet 
> df.write().mode(SaveMode.Append).parquet(jClass.getName()+"_parquet");
> and I get the error Caused by: org.apache.parquet.schema.
> InvalidSchemaException: Cannot write a schema with an empty group: optional group element {}
> Kindly direct how to specify an ArrayList of Strings or Doubles in the Bean passed to the encoders so that the schema generated can be saved in paraquet



--
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