按列的最后一个活动行并将其聚合到另一列

Question

So, I have this kind of transaction data; a cust_id can have one or more acc_no. I want to pull the latest data of cust_id based on the latest data it has by DATE, and aggregate sum it by acc_no.

Sample Data

+-----------+-------------+--------------+-------------------+
|   DATE    |   acc_no    |   cust_id    |  total_balance    |
+-----------+-------------+--------------+-------------------+
| 1/1/2021  |      1      |    A201      |       300         |
| 1/1/2021  |      2      |    A201      |       20          |
| 1/1/2021  |      3      |    A202      |       40          |
| 1/1/2021  |      4      |    A203      |       200         |
| 1/1/2021  |      5      |    A203      |       150         |
| 2/1/2021  |      2      |    A201      |       50          |
| 3/1/2021  |      2      |    A201      |       100         |
| 4/1/2021  |      1      |    A201      |       400         |
| 4/1/2021  |      2      |    A201      |       125         |
| 5/1/2021  |      3      |    A202      |       50          |
| 5/1/2021  |      5      |    A203      |       165         |
+-----------+-------------+--------------+-------------------+

Desired output

+-----------+--------------+-------------------+
|   DATE    |   cust_id    |  total_balance    |
+-----------+--------------+-------------------+
| 4/1/2021  |    A201      |       525         |
| 5/1/2021  |    A202      |       50          |
| 5/1/2021  |    A203      |       365         |
+-----------+--------------+-------------------+

I tried to use:

SELECT *
FROM (
    SELECT  DATE
            ,cust_id
            ,total_balance
            ,ROW_NUMBER() OVER (
            PARTITION BY customer_id ORDER BY DATE DESC
            ) rn
    FROM `a.table`
    ) q
WHERE rn = 1

It will return only one of the acc_no not the aggregate of it

+-----------+--------------+----------------------------------------+
|   DATE    |   cust_id    |            total_balance               |
+-----------+--------------+----------------------------------------+
| 4/1/2021  |    A201      |  sometimes 400 sometimes 125 (random)  |
| 5/1/2021  |    A202      |                  50                    |
| 5/1/2021  |    A203      |                 165                    |
+-----------+--------------+----------------------------------------+

If you have any suggestions on how to create the right query, please let me know. Any suggestions are appreciated.

----Edit----

The answer below by meysam asadi is great, but what if the case of acc_no of the same cust_id is on a different date?

----Edit2----

Adding sample to my questions for more clarity

Answer 1

You can use the following query. I first grouped the records according to cust_id and then joined it to its own table and again grouped the records according to cust_id so that I could use the sum function.

SELECT t2.date,t2.cust_id,SUM(t2.total_balance) AS total_balance               
FROM
(SELECT MAX(date) AS date,cust_id FROM yourTable GROUP BY cust_id) t1 
INNER JOIN yourTable t2 ON t1.cust_id = t2.cust_id AND t1.date = t2.date
GROUP BY t2.cust_id,t2.date