[IMPALA-376] Built-in functions for parsing JSON

• Type: New Feature
• Status: Resolved
• Priority: Minor
• Resolution: Fixed
• Affects Version/s: Product Backlog
• Fix Version/s: Impala 3.1.0
• Component/s: Backend
• Labels: built-in-function
• Environment: All supported environments
• Target Version: Product Backlog
• Assignee: Quanlong Huang
• Reporter: Zoltan Toth-Czifra
• Created: 27/May/13 13:24
• Updated: 20/Dec/18 14:12
• Resolved: 20/Dec/18 14:12

Description

Hi,

Hive comes with some useful built-in UDFs to process JSON objects.

https://cwiki.apache.org/confluence/display/Hive/LanguageManual+UDF

Namely:

• get_json_object
• json_tuple

To make Impala and Hive tables and quieries more interchangable, I am proposing porting these UDFs to be part Impala's built in functions:

http://www.cloudera.com/content/cloudera-content/cloudera-docs/Impala/latest/Installing-and-Using-Impala/ciiu_functions.html

Example

Consider the following table raw_log

action	parameters
search	{"keyword":"hotel"}
visit	{"url":"http://example.com"}

...and the following query:

SELECT get_json_object(event_params, "$.keyword") AS keyword FROM raw_log WHERE action='search';

The query should return the following results:

keyword
hotel

Issue Links

• is blocked by: IMPALA-7364 Upgrade RapidJson to the latest version
• is related to: IMPALA-7609 Support ANSI SQL JSON functions

Sub-Tasks

• Document JSON Udfs: Closed