Question

How to select most recent records if the records are having almost same kind of data…

example:

col1       col2          col3              col4
--------------------------------------------------
123         abc          1.1               12345
123         abc          1.1               123445
1234        dsv          2.0               123
1234        dsv          2.0               1233
12345       dsvw         1.2               1234

The maximum of col4 when it compares with rest of the columns.

See the row1 and row2 data looks similar but we need the latest data based on col4.

Same thing with row3 and row4, data looks similar but we need the latest data based on col4.

And the required output is:

col1    col2   col3   col4
----------------------------
123     abc    1.1    123445
1234    dsv    2.0    1233
12345   dsvw   1.2    1234 

Answer 1

Use:

WITH example AS (
   SELECT t.*,
          ROW_NUMBER() OVER (PARTITION BY t.col1
                                 ORDER BY t.col4 DESC) AS rnk
     FROM YOUR_TABLE t)
SELECT e.*
  FROM example e
 WHERE e.rnk = 1

..or:

SELECT e.*
  FROM (SELECT t.*,
               ROW_NUMBER() OVER (PARTITION BY t.col1
                                      ORDER BY t.col4 DESC) AS rnk
          FROM YOUR_TABLE t) e
 WHERE e.rnk = 1

The CTE doesn’t provide any optimization over the derived table approach.