Select column from another table based on matching condition

Question

I have two tables that have time-series data in the following form. I want to select all of the columns of TABLE1 and an additional column Limit_At_Time which is the value of the Limit column of TABLE2 for the row which has the most recent Date (relative to TABLE1.Date)

TABLE1:

Customer_ID   Date         Amount
1             01/01/2019   5
2             02/05/2019   15

TABLE2:

Customer_ID   Date         Limit
1             12/05/2018   10
1             12/25/2018   20
2             01/05/2019   30
2             03/08/2019   50

Result:

Customer_ID   Date         Amount      Limit_At_Time
1             01/01/2019   5           20
2             02/05/2019   15          30

The closest I have gotten is selecting a previous_date column with this query:

SELECT *, 
(SELECT MAX(Date) FROM TABLE2 t2
   WHERE t2.Date < t1.Date 
   AND t2.Customer_ID = t1.Customer_ID)
as previous_date
FROM TABLE1 AS t1

This gets me the date of the event from TABLE2 that I am interested in for each TABLE1 row, but I need to extract the Limit column value of the row that contains that previous_date .

How can I achieve the result that I want?

Answer 1

You can use your subquery as a JOIN condition between table1 and table2 which will then allow you to get the Limit value from table2 :

SELECT t1.Customer_ID, t1.Date, t1.Amount, t2.Limit
FROM table1 t1
JOIN table2 t2 ON t2.Customer_ID = t1.Customer_ID
              AND t2.Date = (SELECT MAX(Date) FROM table2 t2b
                             WHERE t2b.Date < t1.Date 
                             AND t2b.Customer_ID = t1.Customer_ID)

Output:

Customer_ID Date        Amount  Limit
1           2019-01-01  5       20
2           2019-02-05  15      30

Demo on dbfiddle

Answer 2

I would just use a correlated subquery:

select t1.*,
       (select t2.limit
        from table2 t2
        where t2.date <= t1.date
        order by t2.date desc
        limit 1
       ) as Limit_At_Time
from table1 t1;

Usually in these types of problems, the comparison is <= rather than < , so I used that. Of course, the exact equivalent to your query is < .