-1

i would like to create customer retention.

So i've this table :

id date purchase
1 2020-01 200
2 2020-12 100
2 2020-03 150
3 2020-03 200
1 2020-07 120
1 2020-08 300
3 2020-05 250

and i want this output :

purchase month first purchase month total
0 2020-01 200
6 2020-01 320
7 2020-01 620
0 2020-03 350
4 2020-03 600
11 2020-03 700

"Customers who placed their first order in Jan 2020 spent 200 in month 0 (so in Jan 2020) and 120 (320-200) in month 6 (so in Jul 2020)."

"Customers who placed their first order in Mar 2020 spent 350 in month 0 (so Mar 2020) and 250 (600-350) in month 4 (so May 2020)."

Thanks by advance for your help

1 Answers1

0

You can easily do it with window function and common table expression.

Schema and insert statements:

 create table purchases(id int, date date, purchase int);
 insert into purchases values(1,    '2020-01',  200);
 insert into purchases values(2,    '2020-12',  100);
 insert into purchases values(2,    '2020-03',  150);
 insert into purchases values(3,    '2020-03',  200);
 insert into purchases values(1,    '2020-07',  120);
 insert into purchases values(1,    '2020-08',  300);
 insert into purchases values(3,    '2020-05',  250);     

Query:

with cte as
     ( 
       select id,date,purchase,
       min(date)over(partition by id) FirstPurchaseMonth from purchases
     )
     ,cte2 as
     (
       select substr(date,6,2)-substr(firstpurchasemonth,6,2) Purchasemonth, max(FirstPurchaseMonth)firstpurchasemonth, 
       purchase,sum(purchase)total from cte
       group by firstpurchasemonth,substr(date,6,2)-substr(firstpurchasemonth,6,2) 
     )
     select purchasemonth,firstpurchasemonth,sum(total)over(partition by firstpurchasemonth order by purchasemonth)total
     from cte2 
 

Output:

Purchasemonth firstpurchasemonth total
0 2020-01 200
6 2020-01 320
7 2020-01 620
0 2020-03 350
2 2020-03 600
9 2020-03 700

db<fiddle here