IssueTracker

These 3 queries should behave in a similar way (same results and similar performance), but somehow `NOT IN` and `NOT EXISTS` is behaving way slower right now.

I created a bug to track this performance hit, as it should only be transitory.

(similar situation initially reported by customer https://stackoverflow.com/questions/52550566/bigquery-not-in-comparison-performance-too-slow)

    SELECT tags, COUNT(*) c, ANY_VALUE(b.value)
    FROM `bigquery-public-data.stackoverflow.posts_questions` a
    LEFT JOIN (SELECT x.value FROM UNNEST((
      SELECT APPROX_TOP_COUNT(tags, 10000)
      FROM `bigquery-public-data.stackoverflow.posts_questions`
    )) x ) b
    ON a.tags=b.value
    WHERE b.value IS NULL
    GROUP BY 1
    ORDER BY 2 DESC
    LIMIT 1000
   
    12 seconds, fh-bigquery:US.bquijob_3c0fdf82_1661f6f3dd1

    SELECT tags, COUNT(*) c
    FROM `bigquery-public-data.stackoverflow.posts_questions`
    WHERE tags NOT IN(SELECT x.value FROM UNNEST((
      SELECT APPROX_TOP_COUNT(tags, 10000)
      FROM `bigquery-public-data.stackoverflow.posts_questions`
    )) x)
    GROUP BY 1
    ORDER BY 2 DESC, 1
    LIMIT 1000

    > 350 seconds, fh-bigquery:US.bquijob_766cc8ab_1661f7023bb

    SELECT tags, COUNT(*) c
    FROM `bigquery-public-data.stackoverflow.posts_questions`
    WHERE NOT EXISTS(SELECT x.value FROM UNNEST((
      SELECT APPROX_TOP_COUNT(tags, 10000)
      FROM `bigquery-public-data.stackoverflow.posts_questions`
    )) x WHERE tags=value)
    GROUP BY 1
    ORDER BY 2 DESC, 1
    LIMIT 1000

    > 350 seconds, fh-bigquery:US.bquijob_59a9d1e6_1661f59db40