SQL statement with having, min , max - sql

Question

I have a table:
ID INTEGER NOT NULL, -- AUTOMATIC RECORD'S ID
CUSTOMER_ID INTEGER NOT NULL,
BILING_PERIOD DATE NOT NULL,
DOCUMENT_ID INTEGER NOT NULL,
DATE_CREATED DATE NOT NULL -- WHEN THE DOCUMENT WAS CREATED
I want to select number of documents for customer in biling period,
id for the document that was created first in biling period for customer
and id for the document that was created last in biling period for customer.
All should be sorted by customer and biling period.
I want only biling periods that have more than 1 document for customer.
So when we have for example such data:
ID CUSTOMER_ID BILING_PERIOD DOCUMENT_ID DATE_CREATED
1 5 2020-01-01 123 2020-02-01
2 5 2020-01-01 22 2019-02-01
3 5 2020-01-01 3 2010-02-01
4 99 2020-01-01 458 2021-02-01
5 99 2020-01-01 64 2010-02-01
6 100 2020-01-01 120 2020-02-01
7 99 2019-06-01 452 2019-06-01
8 99 2019-06-01 546 2019-12-01
I want my results looks like that:
CUSTOMER_ID BILING_PERIOD NR_OF_DOC FIRST_DOC_ID LAST_DOC_ID
5 2020-01-01 3 3 123
99 2019-06-01 2 452 546
99 2020-01-01 2 64 458
Myself I can only count number of documents per user and period
SELECT customer_id, biling_period, count(*) as nr_of_doc
FROM T1
GROUP BY customer_id, biling_period
HAVING COUNT() > 1;
CUSTOMER_ID BILING_PERIOD NR_OF_DOC
5 2020-01-01 3
99 2019-06-01 2
99 2020-01-01 2
I do not know hot to get document_id for newest and oldest document.

Answer 1

You can use row_number() and aggregation:
select
customer_id,
billing_period,
count(*),
max(case when rn_asc = 1 then document_id end) fist_doc_id,
max(case when rn_desc = 1 then document_id end) last_doc_id
from (
select
t.*,
row_number() over(
partition by customer_id, billing_period order by date_created
) rn_asc,
row_number() over(
partition by customer_id, billing_period order by date_created desc
) rn_desc
from t1 t
) t
group by customer_id, billing_period
having count(*) > 1
order by customer_id, billing_period
This will wodk properly even if the document ids are not in sequence.
Demo on DB Fiddle:
customer_id | billing_period | count | fist_doc_id | last_doc_id
----------: | :------------- | ----: | ----------: | ----------:
5 | 2020-01-01 | 3 | 3 | 123
99 | 2019-06-01 | 2 | 452 | 546
99 | 2020-01-01 | 2 | 64 | 458

Answer 2

In your sample data, the document ids seem to be assigned in order. If that is the case, you can just use aggregation:
SELECT customer_id, billing_period, count(*) as nr_of_doc,
MIN(document_id), MAX(document_id)
FROM T1
GROUP BY customer_id, billing_period
HAVING COUNT() > 1;