Redshift-将数据分成多行

Question

I have sales data that shows customers that have bought a product. All customer IDs are appended to the same cell as shown below.

How could I split this such that I have one row per cust_id per prod_id

prod_id,cust_id
10001,100,101
10002,102

Expected output:

prod_id,cust_id
10001,100
10001,101
10002,102

Answer 1

Since Redshift supports generate_series now, you can use the following SQL to achieve the desired result

with test(prod_id,cust_id) as (
select 10001,'100,101' union all
select 10002,'102'
),
max_ids as (
select distinct generate_series(1, regexp_count(cust_id,',') +1) n from test
)
select distinct prod_id, split_part(cust_id, ',', n) cust_id
from test, max_ids 
where split_part(cust_id, ',', n) != ''

output is

prod_id cust_id
10001   100
10001   101
10002   102