Question

I have a mysql table to store messages between users. I'd like to calculate the average time a specific user waits to answer to the first message he/she receives from other users

my table is something like this, where subject_id could be the thread or post

|msg_id|sender_id|receiver_id| subject_id |msg           |time               |
--------------------------------------------------------------------------
|   1  |  123    |  456      |     3      |"yadda yadd.."|2016-01-31 00:27:16|
|   2  |  456    |  123      |     3      |"ladida   .." |2016-01-31 00:37:16|
|   3  |  456    |  123      |     3      |"johndo   .." |2016-01-31 01:47:04|
|   4  |  123    |  456      |     3      |"xxxxxx   .." |2016-01-31 02:47:04|
|   5  |  456    |  123      |     3      |"qwerty   .." |2016-01-31 03:47:04|
|   6  |  789    |  456      |     9      |"dadda kadd.."|2016-01-31 00:11:16|
|   7  |  789    |  456      |     9      |"fadda jadd.."|2016-01-31 00:12:16|
|   8  |  456    |  789      |     9      |"fadda jadd.."|2016-01-31 00:13:16|

In this specific case the user 456, received a message from user 123 and responded 10 mins later.

Then he received a message from user 789 and responded in 1 minute.

So the average response time for user 456 is say 5 minutes on average.

UPDATE

to make myself more clear: think a dating app. i want to calculate what's the average response time of a user, so other user will know if she/he is quick or very slow answering messages

Answer 1

Here is a query you can try:

SELECT m1.sender_id, m1.receiver_id,
    AVG(TIMESTAMPDIFF(MINUTE, m1.time, m2.time)) AS DiffInMinutes
FROM messages m1 INNER JOIN messages m2
    ON m1.receiver_id = m2.sender_id AND m1.sender_id = m2.receiver_id AND
        m2.time = (SELECT MIN(time) FROM messages WHERE time > m1.time)
GROUP BY m1.sender_id, m1.receiver_id, m2.sender_id, m2.receiver_id

Output:

╔═══════════╦═════════════╦═══════════════╗
║ sender_id ║ receiver_id ║ DiffInMinutes ║
╠═══════════╬═════════════╬═══════════════╣
║   123     ║     456     ║      35       ║
║   456     ║     123     ║      60       ║
║   789     ║     456     ║      1        ║
╚═══════════╩═════════════╩═══════════════╝

Follow the link below for a running demo:

SQLFiddle

Answer 2

Something like this should work :

SELECT receiver_id, avg(waiting_time)
FROM (
    SELECT m1.receiver_id , min(TIMESTAMPDIFF(MINUTE, m1.time, m2.time)) as `waiting_time`
    FROM messages m1
    JOIN messages m2
        on m1.receiver_id = m2.sender_id
        and m1.sender_id = m2.receiver_id
        and m1.time < m2.time
    WHERE m1.receiver_id = 456
    GROUP BY m1.sender_id, m1.receiver_id
) as tmp;

Result :

╔═════════════╦═══════════════════╗
║ receiver_id ║ avg(waiting_time) ║
╠═════════════╬═══════════════════╣
║     456     ║      5.5000       ║
╚═════════════╩═══════════════════╝

MYSQL calculate average time between messages received and first answer

2 个答案: