Home/ Questions/Q 3990580
In Process

The Archive Base Latest Questions

Editorial Team
  • 0
Asked: 2026-05-20T06:30:45+00:00

Let’s say I’ve got a SQL Server table that logs user activity. Let’s say

  • 0

Let’s say I’ve got a SQL Server table that logs user activity. Let’s say it has user ID, user name, activity date, and activity type columns. I want to print out a list of all user activity, with one row for each month of activity, and a column for each activity type summing up the number of times that activity occurred in that month. I’m trying to do this with the following query:

 SELECT 
    user_id, 
    user_name, 
    CONVERT(VARCHAR(7), activity_date, 120),
    SUM(CASE WHEN activity_type = 'Log In' THEN 1 ELSE 0 END),
    SUM(CASE WHEN activity_type = 'Save Document' THEN 1 ELSE 0 END),
    SUM(CASE WHEN activity_type = 'Create Document' THEN 1 ELSE 0 END)
    FROM UserActivity
    WHERE DATE BETWEEN '11-1-2010 00:00:00' AND '12-31-2010 23:59:59'
GROUP BY user_id, user_name, CONVERT(VARCHAR(7), activity_date, 120)

The problem is, this query is essentially giving me a separate row for each activity–lots and lots of rows, no counting. I think that the problem is with the way I’m doing the dates, because if I change the query to not select the date, I get a table that looks “mostly correct.”

Any thoughts?

Share

Leave an answer

You must login to add an answer.

Need An Account,

1 Answer

  1. Editorial Team

    You can’t have a SUM without a GROUP BY, at least not with other non-aggregates in the SELECT. Do your GROUP BY clause properly.

    SELECT 
user_id, 
user_name, 
CONVERT(VARCHAR(7), activity_date, 120),
SUM(CASE WHEN activity_type = 'Log In' THEN 1 ELSE 0 END),
SUM(CASE WHEN activity_type = 'Save Document' THEN 1 ELSE 0 END),
SUM(CASE WHEN activity_type = 'Create Document' THEN 1 ELSE 0 END)
FROM UserActivity
WHERE DATE BETWEEN '11-1-2010 00:00:00' AND '12-31-2010 23:59:59'
GROUP BY user_id, 
user_name, 
CONVERT(VARCHAR(7), activity_date, 120)

    For what it’s worth, for date ranges, I prefer to use

    WHERE DATE >= '20101101'
  AND DATE <  '20110101'

    I’m sure losing a few records with a timestamp of ’12-31-2010 23:59:59.997′ won’t matter, but it’s just more logically correct to use a < next_date test. And to be pedantic, the format YYYYMMDD is the most robust regardless of regional/language/dateformat settings.

    • 0