[HIVE-2927] Allow escape character in get_json_object

• Type: Improvement
• Status: Open
• Priority: Major
• Resolution: Unresolved
• Affects Version/s: 0.8.1, 0.9.0
• Component/s: Serializers/Deserializers
• Assignee: Unassigned
• Reporter: Sean McNamara
• Created: 05/Apr/12 17:49
• Updated: 23/Oct/18 06:06

Description

Background:
get_json_object extracts json objects from a json string based on a specified path.

Problem:
The current implementation of get_json_object can't see keys with a '.' in them. Our data contains '.' in the keys, so we have to filter our json keys through a streaming script to replace '.' for '_'.

Example:
{{json = {"a":

{"b": 1}

, "c.d": 2}}}

get_json_object(json, "$.a.b") returns: 1
get_json_object(json, "$.c.d") returns: NULL

In the present implementation of get_json_object, c.d is not addressable.

Proposal:
The desired behavior would be to allow the JSON path to be escape-able, like so:

get_json_object(json, '$.c\\\.d') would return: 2