You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@nifi.apache.org by "Matt Burgess (Jira)" <ji...@apache.org> on 2021/11/15 15:53:00 UTC

[jira] [Commented] (NIFI-7862) Allow PutDatabaseRecord to optionally create the target table and/or columns if they don't exist

    [ https://issues.apache.org/jira/browse/NIFI-7862?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17443910#comment-17443910 ] 

Matt Burgess commented on NIFI-7862:
------------------------------------

Per the discussion in NIFI-8101, this Jira will be used to track a new processor UpdateDatabaseTable that works like UpdateHive3Table in the sense that it can add and change columns (and record fields) based on the incoming data. As there are a lot of configuration options it was deemed better to put them in a new processor rather than add the complexity to PutDatabaseRecord, and this is consistent with the UpdateHive3Table -> PutHive3Streaming pattern already in use.

> Allow PutDatabaseRecord to optionally create the target table and/or columns if they don't exist
> ------------------------------------------------------------------------------------------------
>
>                 Key: NIFI-7862
>                 URL: https://issues.apache.org/jira/browse/NIFI-7862
>             Project: Apache NiFi
>          Issue Type: Improvement
>          Components: Extensions
>            Reporter: Matt Burgess
>            Assignee: Matt Burgess
>            Priority: Major
>
> As part of NIFI-6934, the "Database Type" property was added, allowing for DB-specific SQL/DML to be generated. That property could be used with a new true/false property "Create Table" that would create the target table if it didn't already exist. mapping the NiFi Record datatypes to DB-specific column types and generating/executing a DB-specific CREATE TABLE statement. If "Create Table" is set to true and an error occurs while attempting to create the table, the flowfile should be routed to failure or rolled back (depending on the setting of the "Rollback on Failure" property).
> Also it could support some schema migration (adding columns, e.g.) with an additional value in the "Unmatched Field Behavior" property for "Create Column on Unmatched Fields". The use case here is for feature parity with UpdateHiveTable (see NIFI-7989)



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