Count number of times 2 distinct values appear in more than 1 row in SQL Table

Question

I am fairly new to SQL and am having some trouble with the last step in a slightly complicated SQL Query. I want to count how many times in the table, two distinct values appear in more than 1 row.

My specific scenario is, that my table stores Messages/Alerts too and from a system. These alerts are sent to multiple people who are required to respond. I want to count how many of the recipients responded to each Alert.

I have edited down my query to the part i am stuck on, which is this:

SELECT DISTINCT AlertID, count(RecipientID) - count(Distinct RecipientID) as Replies, 
FROM [myDB].[dbo].[Alerts]   

This query shows how many messages each alert has in it, not including the original message to each recipient from the system. The issue with this is that if someone responds twice to a single alert, it is counted as two responses as apposed to 1(which is what i want).

I thought i could do this by counting each DISTINCT AlertID and how many times a DISTINCT recipients ID appears with the AlertID in more than 1 row. It wouldn't matter if someone responded twice, because as long as there is more than one it is counted as a single response. I'm having trouble getting this to working.

My data looks like this:

RecipientID  MsgContents        SentBy       AlertID
12345        Msg1               mySystem     11111
98765        Msg1               MySystem     11111
12345        1st Reply to Msg1  John Doe     11111
12345        2nd Reply to Msg1  John Doe     11111
98765        reply to Msg1      Mike Smith   11111
12345        Msg3               mySystem     33333
12345        Reply to Msg3      John Doe     33333  
12345        Msg2               mySystem     99999

I would like the output of the query to be:

 AlertID        Replies
 11111          2
 33333          1
 99999          0   

I thought this answer would help, but couldn't get it to work for me. I also would like to keep it in one query if possible, as the system will be dealing with large amounts of data. If this is not the easiest way to do it, i am open to all suggestions.

Any help would be greatly appreciated.

Answer 1

You can do this using a CASE expression inside COUNT .

SELECT
    AlertID,
    Replies = COUNT(DISTINCT CASE WHEN SentBy <> 'mySystem' THEN RecepientID END)
FROM alerts
GROUP BY AlertID

ONLINE DEMO

Answer 2

First, don't use select distinct when you should be using group by .

Then this produces the results you seem to want:

SELECT AlertID, (count(Distinct RecipientID) - 1) as Replies
FROM [myDB].[dbo].[Alerts]   
GROUP BY AlertId;

You might actually want:

SELECT AlertID, count(Distinct case when sentBy <> 'mySystem' then RecipientID) as Replies 
FROM [myDB].[dbo].[Alerts]   
GROUP BY AlertId;

Answer 3

select AlertId, count(RecipientId) as Replies group by AlertId 
from (select * from Alerts where SentBy <> 'mySystem') 

--The first line gets records for each alert

-- The second line gets records for actual recipients as opposed to system records