『Treasure Data でアクセスログ分析の限界に挑む』その③ 〜アクセスに基づいたユーザーセグメントの作成 月次編（平日/土日）〜 ref: https://qiita.com/doryokujin/items/89392c4cc0bc00c71629

file0.sql

TD_TIME_FORMAT(time,'u','JST') # 1(月曜日) 〜 7（日曜日）

file1.sql

SELECT         
  SUM(IF(week_num     IN ('6','7'), 1,0)) AS freq_weekend_total,
  SUM(IF(week_num NOT IN ('6','7'), 1,0)) AS freq_weekday_total
FROM
(
  SELECT TD_DATE_TRUNC('day',time,'JST') AS access_day, TD_TIME_FORMAT(time,'u','JST') AS week_num
  FROM pageviews    
  WHERE TD_INTERVAL(time, '-1M', 'JST')
  GROUP BY TD_DATE_TRUNC('day',time,'JST'), TD_TIME_FORMAT(time,'u','JST')
)

file2.sql

/* TD_SCHEDULED_TIME() = '2018-11-23 11:11:00' */
/**/
SELECT target_month, 
  CASE 
    WHEN 70 <= ratio_freq_weekend OR 70 <= ratio_pv_weekend THEN 'weekend'
    WHEN ratio_freq_weekend <= 30 OR ratio_pv_weekend <= 30 THEN 'weekday'
    WHEN ratio_freq_weekend IS NULL THEN 'non_active'
    ELSE 'allday'
  END AS segment,
  COUNT(1) AS cnt
FROM
(
/**/
  SELECT
    TD_TIME_FORMAT(TD_DATE_TRUNC('month', TD_DATE_TRUNC('month', TD_SCHEDULED_TIME(),'JST')-1, 'JST'),'yyyy-MM-dd','JST') AS target_month,
    IF(past_month.td_client_id IS NOT NULL, past_month.td_client_id, past_month_ago.td_client_id ) AS td_client_id,
    freq_weekend, freq_weekend_total,
    freq_weekday, freq_weekday_total,
    ROUND(1.0*freq_weekend/freq_weekend_total,2)*100 AS freq_weekend_ratio,
    ROUND(1.0*freq_weekday/freq_weekday_total,2)*100 AS freq_weekday_ratio,
    ROUND(
      ( 1.0*freq_weekend/freq_weekend_total ) / ( 1.0*freq_weekend/freq_weekend_total + 1.0*freq_weekday/freq_weekday_total )
    ,2)*100 AS ratio_freq_weekend,
    pv_weekend, pv_weekday,
    ROUND(1.0*pv_weekend/freq_weekend_total,2)*100 AS pv_weekend_avg,
    ROUND(1.0*pv_weekday/freq_weekday_total,2)*100 AS pv_weekday_avg,
    ROUND(
      ( 1.0*pv_weekend/freq_weekend_total ) / ( 1.0*pv_weekend/freq_weekend_total + 1.0*pv_weekday/freq_weekday_total )
    ,2)*100 AS ratio_pv_weekend,
    freq_weekend_total, freq_weekday_total,
    min_time, max_time
  FROM
  (
    SELECT td_client_id, freq_weekend_total, freq_weekday_total, freq_weekend, freq_weekday, pv_weekend, pv_weekday, min_time, max_time
    FROM
    (
      SELECT
        td_client_id,
        SUM(IF(week_num     IN ('6','7'),  1,0)) AS freq_weekend,
        SUM(IF(week_num NOT IN ('6','7'),  1,0)) AS freq_weekday,
        SUM(IF(week_num     IN ('6','7'), pv,0)) AS  pv_weekend,
        SUM(IF(week_num NOT IN ('6','7'), pv,0)) AS  pv_weekday,
        TD_TIME_FORMAT(MIN(min_time),'yyyy-MM-dd','JST') AS min_time, TD_TIME_FORMAT(MAX(max_time),'yyyy-MM-dd','JST') AS max_time
      FROM
      (
        SELECT td_client_id, TD_DATE_TRUNC('day',time,'JST') AS access_day, TD_TIME_FORMAT(time,'u','JST') AS week_num, COUNT(1) AS pv, MIN(time) AS min_time, MAX(time) AS max_time
        FROM pageviews    
        WHERE TD_INTERVAL(time, '-1M', 'JST')
        GROUP BY td_client_id, TD_DATE_TRUNC('day',time,'JST'), TD_TIME_FORMAT(time,'u','JST')
      ) 
      GROUP BY td_client_id
    ),
    (
      SELECT         
        SUM(IF(week_num     IN ('6','7'), 1,0)) AS freq_weekend_total,
        SUM(IF(week_num NOT IN ('6','7'), 1,0)) AS freq_weekday_total
      FROM
      (
        SELECT TD_DATE_TRUNC('day',time,'JST') AS access_day, TD_TIME_FORMAT(time,'u','JST') AS week_num
        FROM pageviews    
        WHERE TD_INTERVAL(time, '-1M', 'JST')
        GROUP BY TD_DATE_TRUNC('day',time,'JST'), TD_TIME_FORMAT(time,'u','JST')
      )
    )
  ) past_month
  FULL OUTER JOIN
  (
    SELECT td_client_id
    FROM pageviews    
    WHERE TD_INTERVAL(time, '-10y/-1M', 'JST') /* 1ヶ月前より過去 */
    GROUP BY td_client_id
  ) past_month_ago
  ON past_month.td_client_id = past_month_ago.td_client_id
  /* 前月分のみのユーザーのみ取得で良いのならば past_month_ago サブクエリは不要 */
/**/
)
GROUP BY target_month,
  CASE 
    WHEN 70 <= ratio_freq_weekend OR 70 <= ratio_pv_weekend THEN 'weekend'
    WHEN ratio_freq_weekend <= 30 OR ratio_pv_weekend <= 30 THEN 'weekday'
    WHEN ratio_freq_weekend IS NULL THEN 'non_active'
    ELSE 'allday'
  END
/**/