How to calculate date difference for group data

Question

I would like to count the date-time difference, when I group the data.

Lets have a look on table content example :

id | session_id | created
1  | 101        | 1/10/2010 9:30:10
2  | 102        | 1/10/2010 9:31:10
3  | 101        | 1/10/2010 9:32:10
4  | 103        | 1/10/2010 9:35:10
5  | 102        | 1/10/2010 9:38:10
6  | 103        | 1/10/2010 9:39:10

Output should be as follow :

session_id | time_count
101        | 2 (minutes)
102        | 7 (minutes)
103        | 4 (minutes)

So here i want to calculate time difference of 'created' field for data which is group by session_id.

Any help would be great. Thanks in advance :)

My Situation :

I have a table with following fields :

id, session_id, platform, event, created

there are 2 event (Start,End)

Now if there are records of Start and End for same session_id then i want the time takes to complete this session.

But if i have 2 Start session, but only 1 End session then i do not want the time difference for 2nd Session because it did not end

I agree that you need to describe the situation more. However please add the DDL statements with INSERT statements as well. It will help to give you the right answer — Dmitrij Kultasev, May 03 '18 at 05:56

score 0 · Answer 1 · answered May 03 '18 at 05:54

0

SELECT session_id,
DATEDIFF(MINUTE, MAX(created), MIN(created)) AS diff
FROM table
GROUP BY session_id

answered May 03 '18 at 05:54

Dmitrij Kultasev

5,447
5
44
88

DineshDB · Answer 2 · 2018-05-03T10:08:01.717

Try this:

Table Schema:

CREATE TABLE A(id INT, session_id INT,Event VARCHAR(20), created DATETIME);

INSERT INTO A VALUES(1, 101,'Start','2010/10/01 9:30:10')
,(2, 102,'Start'  , '2010/10/01 9:31:10')
,(3, 101,'End'    , '2010/10/01 9:32:10')
,(4, 103,'Start'  , '2010/10/01 9:35:10')
,(5, 102,'End'    , '2010/10/01 9:38:10')
,(6, 103,'End'    , '2010/10/01 9:39:10')
,(7, 101,'Start'  , '2010/10/01 9:39:10')
,(8, 102,'Start'  , '2010/10/01 9:39:10')
,(9, 101,'End'    , '2010/10/01 9:55:10');

Query:

SELECT D.session_id
    ,TIMESTAMPDIFF(MINUTE,MIN(D.created), MAX(D.created)) time_count
FROM(
  SELECT a.id,a.session_id,a.created
    ,ROUND(count(*)/2,0) as RN 
    ,count(*) as row_number1 
  FROM a AS a
  JOIN a AS b ON a.session_id = b.session_id AND a.id >= b.id
  GROUP BY a.id,a.session_id,a.created
  ORDER BY 2,3
    )D
GROUP BY D.session_id,D.RN
HAVING COUNT(1)>1

Output:

session_id  time_count
101         2
102         7
103         4
101         16

Fiddle:

Check the Output in the #SQL Fiddle

Abdul Rahmat · Answer 3 · 2018-05-03T05:55:45.037

-1

You can try it's sintax :

    WITH cte AS (
    SELECT
        session_id,
        DATEDIFF(minute, created, LAG(created, 1)) AS diff,
        ROW_NUMBER() OVER (PARTITION BY session_id ORDER BY created) rn
    FROM yourTable
)

SELECT

        session_id,
        diff AS time_count
    FROM cte
    WHERE rn % 2 = 0

;

edited May 03 '18 at 05:55

answered May 03 '18 at 05:54

Abdul Rahmat

196
9

1

Would you like to improve this code-only answer by adding an explanation? How does this work to solve OPs problem? You might also want to highlight the differences to other, similar, older answers. – Yunnosch May 03 '18 at 06:04

How to calculate date difference for group data

3 Answers3