Get most frequent usage of Gradle Internal APIs from BigQuery GitHub Data

gradle-internal-api-usage.sql

SELECT
  class,
  COUNT(*) count
FROM (
  SELECT
    REGEXP_REPLACE(line, r'import |;', '') AS class,
    c.id
  FROM (
    SELECT
      SPLIT(c.content, '\n') line,
      c.id
    FROM
      [gradle.gradle_imports] c
    JOIN
      [bigquery-public-data:github_repos.files] f
    ON
      c.id = f.id
    WHERE
      f.path LIKE '%.java'
        OR f.path LIKE '%.groovy'
        OR f.path LIKE '%.gradle'
    HAVING
      LEFT(line, 17)='import org.gradle' )
  GROUP BY
    class,
    c.id )
WHERE
  NOT REGEXP_MATCH(class, r'org.gradle.(api|authentication|buildinit|caching|external|javadoc|ide|ivy|jvm|language|maven|nativeplatform|platform|play|plugin.devel|plugin.repository|plugin.use|plugins|process|testfixtures|testing.jacoco|tooling|model|testkit|testing).')
  OR class CONTAINS '.internal'
GROUP BY
  1
ORDER BY
  count DESC
LIMIT
  1000;

populate-gradle-imports.sql

-- Save to gradle_imports table
SELECT
  id,
  content
FROM
  [bigquery-public-data:github_repos.contents]
WHERE
  content CONTAINS 'import org.gradle'

-- This is why BigQuery is awesome —> Query complete (21.0s elapsed, 1.94 TB processed)

results-20170614.csv