Revised JSON SerDe Code
Fetching latest commit…
Cannot retrieve the latest commit at this time
Branched from the version of the Google Code project as of July 5. Changes to the Google Code Project after July 5, 2011 are not reflected here. The main goals of this branch were as follows: 1. Incorporate JSON Path as a standardized way of accessing JSON values. 2. Provide a method for users to specify the exact JSON elements which they want to import into Hive. In particular, we added a way of mapping JSON Path expressions to column names (similar to the way that Hive and HBase currently interact). A concrete example is provided below: CREATE EXTERNAL TABLE IF NOT EXISTS response_log_example ( request_id string, ) ROW FORMAT SERDE "org.apache.hadoop.hive.contrib.serde2.JsonSerde" WITH SERDEPROPERTIES ( "request_id"="$.search_result.requestId" ); 3. Better handling of nested JSON structures via JSON Path. The code still does not make any attempt to handle arrays or maps and explicitly forbids any JSON Path expressions which are not "definite."