BigQuery query doesn't work with UNNEST()

Question

I'm trying to search on StackOverflow data through BigQuery by letting this query match a string pattern on answers and by filtering relevant question answers by tags.

WITH question_answers_join AS (
  SELECT *
  FROM (
    SELECT id, creation_date, title
      , (SELECT AS STRUCT body b
         FROM `bigquery-public-data.stackoverflow.posts_answers` 
         WHERE a.id=parent_id
      ) answers
      , SPLIT(tags, '|') tags
    FROM `bigquery-public-data.stackoverflow.posts_questions` a
  )
)SELECT * 
FROM question_answers_join
WHERE 'google-bigquery' IN UNNEST(tags)
AND REGEXP_CONTAINS(answers.b, r"hello")
ORDER BY RAND()
LIMIT 100

however, I get this error:

Scalar subquery produced more than one element

what is it referring to? How can I fix this?

Answer 1

Below is for BigQuery Standard SQL

#standardSQL
WITH question_answers_join AS (
  SELECT *
  FROM (
    SELECT id, creation_date, title
      , ARRAY(SELECT body                             /* this line was the reason for error */
         FROM `bigquery-public-data.stackoverflow.posts_answers` 
         WHERE a.id=parent_id
      ) answers
      , SPLIT(tags, '|') tags
    FROM `bigquery-public-data.stackoverflow.posts_questions` a
  )
)
SELECT *
FROM question_answers_join 
WHERE 'google-bigquery' IN UNNEST(tags)
AND EXISTS (
  SELECT 1 
  FROM UNNEST(answers) answer 
  WHERE REGEXP_CONTAINS(answer, r"hello")
)
ORDER BY RAND()
LIMIT 100    

I think, it is easy to just compare above with your original query to see the differences (hint: there are just two of them). First difference is the actual reason for the error you saw. Second difference is to reflect changes introduced by first one