Hi community, Some of our users store JSON/FlatBuffer data in Impala tables. They wish to perform flexible queries on these data. Common Expression Language (CEL) is a simple expression language built on top of protocol buffer types. It would be nice if there are some built-in functions to support query JSON/FlatBuffer objects with CEL. For example:
- bigint cel_json_eval_int(string input_json, string cel_expression): queries the input_json with cel_expression, returns a bigint type. - bigint cel_json_eval_bool(string input_json, string cel_expression): queries the input_json with cel_expression, returns a boolean type. - string cel_json_eval_string(string input_json, string cel_expression): queries the input_json with cel_expression, returns a string type. - ... With the support of these functions and CEL evaluation, users can write the following query to: - SELECT COUNT(*) FROM events WHERE cel_json_eval_bool(products_detail, "items.size()>=3") AND event_name = 'SubmitOrder' The cel_expression can be any value as long as it follows the CEL language definition (https://github.com/google/cel-spec/blob/master/doc/langdef.md). As a reference, google has a CEL implementation here: https://github.com/google/cel-cpp Best regards, Jian Zhang.
