数据链路分析报告

SELECT
  CASE WHEN chart_name = '微博热搜榜' THEN '热搜榜'
       WHEN chart_name = '微博文娱榜' THEN '文娱榜'
       ELSE chart_name END AS rank_name,
  COUNT(*) as row_cnt,
  COUNT(DISTINCT keywords) as keyword_cnt
FROM external_data.dwd_data_hunter_noteblock_hot_search_board_inc_hi
WHERE dt = '2026-03-26' AND platform_type = 'WEIBO'
GROUP BY 1
ORDER BY 1;

SELECT
  COUNT(*) as row_cnt,
  COUNT(DISTINCT GET_JSON_OBJECT(body, '$.keyword')) as keyword_cnt
FROM external_data.ods_crawler_return_topic
WHERE dt = '2026-03-26'
  AND tag = 'ClairvoyanceWeiBoAiKeywordSearchInfo';

SELECT
  COUNT(DISTINCT t1.keywords) as t1_total_kw,
  COUNT(DISTINCT t2.keyword) as t1_match_t2
FROM (
  SELECT DISTINCT keywords
  FROM external_data.dwd_data_hunter_noteblock_hot_search_board_inc_hi
  WHERE dt = '2026-03-26' AND platform_type = 'WEIBO'
) t1
LEFT JOIN (
  SELECT DISTINCT GET_JSON_OBJECT(body, '$.keyword') as keyword
  FROM external_data.ods_crawler_return_topic
  WHERE dt = '2026-03-26'
    AND tag = 'ClairvoyanceWeiBoAiKeywordSearchInfo'
) t2 ON t1.keywords = t2.keyword;

SELECT
  COUNT(*) as t12_kw,
  SUM(CASE WHEN has_post > 0 THEN 1 ELSE 0 END) as t12_match_t3
FROM (
  SELECT t12.keyword, COUNT(DISTINCT t3.post_id) as has_post
  FROM (
    SELECT DISTINCT t2.keyword
    FROM (...) t1
    INNER JOIN (...) t2 ON t1.keywords = t2.keyword
  ) t12
  LEFT JOIN external_data.dwd_data_hunter_wb_post_inc_hi t3
    ON t3.dt = '2026-03-26'
    AND FIND_IN_SET(t12.keyword, t3.keyword_set) > 0
  GROUP BY t12.keyword
) sub;

SELECT
  CASE WHEN t2.keyword IS NOT NULL THEN '表2有记录'
       ELSE '表2无记录' END as in_t2,
  COUNT(DISTINCT t1.keywords) as kw_cnt
FROM (...) t1
LEFT JOIN (...) t2_match ON t1.keywords = t2_match.keyword
LEFT JOIN (...) t2 ON t1.keywords = t2.keyword
WHERE t2_match.keyword IS NULL
GROUP BY 1;

SELECT
  CASE
    WHEN GET_JSON_OBJECT(GET_JSON_OBJECT(body, '$.result'), '$.msg') IS NULL THEN 'msg为NULL'
    WHEN ... = '' THEN 'msg为空串'
    WHEN LENGTH(...) < 10 THEN 'msg极短(<10字)'
    ELSE 'msg正常'
  END as msg_status,
  ...
FROM external_data.ods_crawler_return_topic
WHERE dt = '2026-03-26'
  AND tag = 'ClairvoyanceWeiBoAiKeywordSearchInfo'
GROUP BY 1, 2;

SELECT
  CASE WHEN t1.keywords IS NOT NULL THEN '匹配微博榜单'
       ELSE '非微博来源' END as source,
  COUNT(DISTINCT t2.keyword) as keyword_cnt
FROM (...) t2
LEFT JOIN (...) t1 ON t2.keyword = t1.keywords
GROUP BY 1;

SELECT
  CASE WHEN chart_name = '微博热搜榜' THEN '热搜榜'
       WHEN chart_name = '微博文娱榜' THEN '文娱榜'
       ELSE chart_name END AS rank_name,
  hh,
  COUNT(*) as row_cnt
FROM external_data.dwd_data_hunter_noteblock_hot_search_board_inc_hi
WHERE dt = '2026-03-26' AND platform_type = 'WEIBO'
GROUP BY 1, 2
ORDER BY 1, 2;

SELECT
  keywords as keyword,
  GROUP_CONCAT(DISTINCT ..., '|') as rank_names,
  COUNT(DISTINCT ...) as chart_cnt,
  MIN(CAST(ranking AS INT)) as min_rank,
  MAX(search_nums) as max_hot_value,
  MIN(update_time) as first_time,
  COUNT(*) as record_cnt
FROM external_data.dwd_data_hunter_noteblock_hot_search_board_inc_hi
WHERE dt = '2026-03-26' AND platform_type = 'WEIBO'
GROUP BY keywords;

SELECT rank_name,
  CASE WHEN min_rank <= 10 THEN 'Top10'
       WHEN min_rank <= 20 THEN 'Top11-20'
       WHEN min_rank <= 30 THEN 'Top21-30'
       WHEN min_rank <= 40 THEN 'Top31-40'
       ELSE 'Top41+' END AS rank_segment,
  COUNT(*) AS total_kw,
  SUM(CASE WHEN t2_hit = 1 THEN 1 ELSE 0 END) AS hit_kw,
  ROUND(...) AS hit_rate
FROM (
  SELECT t1.keywords, ...,
    MIN(CAST(t1.ranking AS INT)) AS min_rank,
    MAX(CASE WHEN t2.keyword IS NOT NULL THEN 1 ELSE 0 END) AS t2_hit
  FROM ... t1 LEFT JOIN ... t2
  WHERE t1.chart_name IN ('微博热搜榜', '微博文娱榜')
  GROUP BY ...
) sub
GROUP BY rank_name, rank_segment
ORDER BY rank_name, MIN(min_rank);

-- 与 A10 结构类似，将 MIN(ranking) 替换为 MIN(CAST(hh AS INT))
-- 外层 CASE 按时段分组：凌晨(00-05) / 上午(06-11) / 下午(12-17) / 晚间(18-23)
SELECT rank_name, time_segment, total_kw, hit_kw, hit_rate
FROM (...同 A10 模式...)
ORDER BY rank_name, MIN(first_hour);

-- 与 A10 结构类似，将 MIN(ranking) 替换为 COUNT(DISTINCT hh)
-- 外层 CASE 按持续时长分组：仅1次 / 2-3次 / 4-6次 / 7次+
SELECT rank_name, duration_segment, total_kw, hit_kw, hit_rate
FROM (...同 A10 模式...)
ORDER BY rank_name, MIN(appear_cnt);