SQL 实现月度留存率/复购率

s q l实现用户的月度留存率/复购率

准备阶段：建立一个中间表 用户ID，用户复购或者访问时间即可。

注意点：注意去重

最好，跑出结果以后自己验证一下。

DROP TABLE IF EXISTS bigdata2c.temp_shoe_daidai_1118_04 ;
CREATE TABLE IF NOT EXISTS bigdata2c.temp_shoe_daidai_1118_04 AS
SELECT  distinct a.order_month
                    ,count(distinct a.user_id_month_0)  
                    ,count(distinct b.client_str)
                    ,count(distinct c.client_str) 
                    ,count(distinct d.client_str) 
                    ,count(distinct f.client_str) 
                    ,count(distinct g.client_str)
                    ,count(distinct h.client_str) 
                    ,count(distinct i.client_str) 
                    ,count(distinct j.client_str) 
                    ,count(distinct k.client_str) 
                    ,count(distinct l.client_str) 
                    ,count(distinct m.client_str) 
                    ,count(distinct n.client_str) FROM    (
                       SELECT  distinct  SUBSTR(create_time,1,6) as order_month
                                ,client_str AS user_id_month_0
                        FROM    bigdata2c.temp_shoe_daidai_1118_01
                        where client_str is not NULL 
                        and create_time is not NULL 
                        and SUBSTR(create_time,1,6) >=201910
                        and SUBSTR(create_time,1,6) <=202010
                    ) a
            LEFT JOIN bigdata2c.temp_shoe_daidai_1118_02 b
            ON      DATEDIFF( to_date(b.order_month,'yyyymm'),to_date(a.order_month,'yyyymm'),'mm') = 1
            AND     a.user_id_month_0 = b.client_str 
            LEFT JOIN    bigdata2c.temp_shoe_daidai_1118_02 c
            ON      DATEDIFF( to_date(c.order_month,'yyyymm'),to_date(a.order_month,'yyyymm'),'mm') = 2
            AND     a.user_id_month_0 = c.client_str
            LEFT JOIN bigdata2c.temp_shoe_daidai_1118_02 d
            ON      DATEDIFF( to_date(d.order_month,'yyyymm'),to_date(a.order_month,'yyyymm'),'mm') = 3
            AND     a.user_id_month_0 = d.client_str LEFT
            JOIN    bigdata2c.temp_shoe_daidai_1118_02 f
            ON      DATEDIFF( to_date(f.order_month,'yyyymm'),to_date(a.order_month,'yyyymm'),'mm') = 4
            AND     a.user_id_month_0 = f.client_str
            LEFT JOIN bigdata2c.temp_shoe_daidai_1118_02 g
            ON      DATEDIFF( to_date(g.order_month,'yyyymm'),to_date(a.order_month,'yyyymm'),'mm') = 5
            AND     a.user_id_month_0 = g.client_str LEFT
            JOIN    bigdata2c.temp_shoe_daidai_1118_02 h
            ON      DATEDIFF( to_date(h.order_month,'yyyymm'),to_date(a.order_month,'yyyymm'),'mm') = 6
            AND     a.user_id_month_0 = h.client_str
            LEFT JOIN bigdata2c.temp_shoe_daidai_1118_02 i
            ON      DATEDIFF( to_date(i.order_month,'yyyymm'),to_date(a.order_month,'yyyymm'),'mm') = 7
            AND     a.user_id_month_0 = i.client_str LEFT
            JOIN    bigdata2c.temp_shoe_daidai_1118_02 j
            ON      DATEDIFF( to_date(j.order_month,'yyyymm'),to_date(a.order_month,'yyyymm'),'mm') = 8
            AND     a.user_id_month_0 = j.client_str
            LEFT JOIN bigdata2c.temp_shoe_daidai_1118_02 k
            ON      DATEDIFF( to_date(k.order_month,'yyyymm'),to_date(a.order_month,'yyyymm'),'mm') = 9
            AND     a.user_id_month_0 = k.client_str LEFT
            JOIN    bigdata2c.temp_shoe_daidai_1118_02 l
            ON      DATEDIFF( to_date(l.order_month,'yyyymm'),to_date(a.order_month,'yyyymm'),'mm') = 10
            AND     a.user_id_month_0 = l.client_str
            LEFT JOIN bigdata2c.temp_shoe_daidai_1118_02 m
            ON      DATEDIFF( to_date(m.order_month,'yyyymm'),to_date(a.order_month,'yyyymm'),'mm') = 11
            AND     a.user_id_month_0 = m.client_str LEFT
            JOIN    bigdata2c.temp_shoe_daidai_1118_02 n
            ON      DATEDIFF( to_date(n.order_month,'yyyymm'),to_date(a.order_month,'yyyymm'),'mm') = 12
            AND     a.user_id_month_0 = n.client_str
group by a.order_month
;