[
https://issues.apache.org/jira/browse/BEAM-12911?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Brian Hulette updated BEAM-12911:
---------------------------------
Priority: P3 (was: P2)
> Errors hydrating beam:coder:row:v1 should indicate the problematic field
> ------------------------------------------------------------------------
>
> Key: BEAM-12911
> URL: https://issues.apache.org/jira/browse/BEAM-12911
> Project: Beam
> Issue Type: Improvement
> Components: sdk-java-core, sdk-py-core
> Reporter: Brian Hulette
> Assignee: Brian Hulette
> Priority: P3
>
> When creating beam:coder:row:v1 instances, both the Java SDK and Python SDK
> can fall back to use Python- or Java- native coders (e.g.
> beam:logical:javasdk:v1). This can be problematic in multi-language
> pipelines, where these types are almost never understood by the other SDK(s).
> This leads to obtuse, non-actionable error messages, like:
> {code}
> ValueError: No logical type registered for URN 'beam:logical_type:javasdk:v1'
> {code}
> We should modify coder hydration logic in both Java and Python to identify
> the field that has an issue, so the user may have a change to understand and
> resolve it.
--
This message was sent by Atlassian Jira
(v8.3.4#803005)
