[HIVE-18295] Add ability to ignore invalid values in JSON SerDe

• Type: Improvement
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Component/s: HCatalog
• Labels: pull-request-available
• Assignee: Matthew Knox
• Reporter: Matthew Knox
• Created: 18/Dec/17 21:41
• Updated: 24/Jun/20 00:26

Description

It would be nice to be able to configure the JSON SerDe to ignore invalid values while parsing JSON.

In our case our raw JSON data is ingested from multiple sources, some of which unreliably sanitize the data. Our current practice is to cleanse the data after ingestion, but that can lead to other issues as well. Having the ability to simply default to NULL if a value can not be parsed would be immensely helpful to us.

Issue Links

• is related to: HIVE-18785 Make JSON SerDe First-Class SerDe
• links to: GitHub Pull Request #278
• links to: GitHub Pull Request #278