如何在大查询中将正则表达式与JSON_EXTRACT一起使用

Question

I have a JSON field on big query table and currently i'm using the following method to do the extraction from a id element (for example):

coalesce(
   nullif(JSON_EXTRACT(e.event_payload, 'content_id'), ''),
   nullif(JSON_EXTRACT(e.event_payload, 'cid'), ''),
   nullif(JSON_EXTRACT(e.event_payload, 'c_id'), ''),
   ...
  ) AS content_id,

I don't have a pattern on this JSON's fields... Is possible use REGEX with JSON_EXTRACT on big query like this?

JSON_EXTRACT(e.event_payload, "(content_id|cid|c_id)") as content_id

Answer 1

You can replace the possible field names and then perform extraction:

SELECT
  JSON_EXTRACT(
    REGEXP_REPLACE(e.event_payload, r'"c\_?id"', '"content_id"'),
    '$.content_id') as content_id
FROM dataset.table

As a self-contained example:

WITH T AS (
  SELECT '{"cid": {"a": 1}}' AS event_payload UNION ALL
  SELECT '{"content_id": {"b": 2}}' UNION ALL
  SELECT '{"c_id": {"c": 3}}'
)
SELECT
  JSON_EXTRACT(
    REGEXP_REPLACE(e.event_payload, r'"c\_?id"', '"content_id"'),
    '$.content_id') as content_id
FROM T AS e