xiangfu0 commented on pull request #7644: URL: https://github.com/apache/pinot/pull/7644#issuecomment-953293632
> > > > Can you provide some reasons a to why we are adding this config? Isnt it easier for the client to just ignore the exception? Is there an ask (issue) for such a thing? > > > > Also, if we _must_ add a config, can it be on a per-table level instead of across all tables on a broker? > > > > > > > > > The purpose is to have a way to keep compatibility but just track if any existing user is accidentally on the old behavior. > > > Once finished the investigation, we should just deprecate this config(maybe after 0.9.0 release) and always throw exception. > > > > > > Maybe I am missing something. I am sure the user code had some exception or other to handle before, so why bother with compatibility here? If they send an invalid query, nothing can be done anyway > > Currently there is no way to find out whether invalid column is being used in the query; an empty response instead of an exception will be returned once the invalid column is included in the query. The purpose of having a metric here is to identify such use cases and then force the users to properly handle such cases. > > > > Can you provide some reasons a to why we are adding this config? Isnt it easier for the client to just ignore the exception? Is there an ask (issue) for such a thing? > > > > Also, if we _must_ add a config, can it be on a per-table level instead of across all tables on a broker? > > > > > > > > > The purpose is to have a way to keep compatibility but just track if any existing user is accidentally on the old behavior. > > > Once finished the investigation, we should just deprecate this config(maybe after 0.9.0 release) and always throw exception. > > > > > > Maybe I am missing something. I am sure the user code had some exception or other to handle before, so why bother with compatibility here? If they send an invalid query, nothing can be done anyway > > Currently there is no way to find out whether invalid column is being used in the query; an empty response instead of an exception will be returned once the invalid column is included in the query. The purpose of having a metric here is to identify such use cases and then force the users to properly handle such cases. I think @mcvsubbu's point is there is no need to identify those cases and they should expect the behavior changes as it helps them discover the user side errors. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
