[jira] [Updated] (HUDI-3596) Understand why NULLSchemaProvider is used for empty batch in InputBatch instead of empty schema

["sivabalan narayanan (Jira)" <ji...@apache.org>]

     [ https://issues.apache.org/jira/browse/HUDI-3596?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

sivabalan narayanan updated HUDI-3596:
--------------------------------------
    Fix Version/s: 0.11.0

> Understand why NULLSchemaProvider is used for empty batch in InputBatch instead of empty schema
> -----------------------------------------------------------------------------------------------
>
>                 Key: HUDI-3596
>                 URL: https://issues.apache.org/jira/browse/HUDI-3596
>             Project: Apache Hudi
>          Issue Type: Task
>          Components: deltastreamer
>            Reporter: sivabalan narayanan
>            Priority: Major
>             Fix For: 0.11.0
>
>
> In [InputBatch|https://github.com/apache/hudi/blob/548000b0d635067acf7574c4cf5122759e79b52b/hudi-utilities/src/main/java/org/apache/hudi/utilities/sources/InputBatch.java#L57] we use NullSchemaProvider when batch is empty. this will return null schema (
> Schema.create(Schema.Type.NULL)). We need to understand why can't we return an empty schema here (Scehma.create(Collections.emptyList()). Thats the schema for an empty spark dataframe and we should stick to that. 
>  
>  
>  



--
This message was sent by Atlassian Jira
(v8.20.1#820001)