SQL query: please help weird output on SQL query. (Counting each client latest status) - sql

Question

I used to ask this question in this SO and already got the solution, however, that sql query produced only partial correct result. I have tried to figure it out but it seem to complicate for me to understand. So would you please kindly help me. Thank you.
for table tbl_sActivity
act_id| Client_id | act_status| user_id | act_date
1 | 7 | warm | 1 | 19/7/12
2 | 7 | dealed | 1 | 30/7/12 <- lastest status of client(7)
3 | 8 | hot | 1 | 6/8/12 <- lastest status of client(8)
4 | 5 | cold | 22 | 7/8/12 <- lastest status of client(5)
5 | 6 | cold | 1 | 16/7/12
6 | 6 | warm | 1 | 18/7/12
7 | 6 | dealed | 1 | 7/8/12 <- lastest status of client(6)
8 | 9 | warm | 26 | 2/8/12
9 | 10 | warm | 26 | 2/8/12
10 | 9 | hot | 26 | 4/8/12 <- lastest status of client(9)
11 | 10 | hot | 26 | 4/8/12
12 | 10 | dealed | 26 | 10/8/12 <- lastest status of client(10)
13 | 13 | dealed | 26 | 8/8/12 <- lastest status of client(13)
14 | 11 | hot | 25 | 8/8/12
15 | 11 | dealed | 25 | 14/8/12 <- lastest status of client(11)
I want to produce the User Progressive Report that shows the latest progress of how each user follow up his/her clients.
Therefore this report must show number of each user's clients group by each of their latest status.
The correct output should be looked like this..
user_id | act_status | Count(act_status)
1 | dealed | 2
1 | hot | 1
22 | cold | 1
25 | dealed | 1
26 | hot | 1
26 | dealed | 2
The SQL query that I had is below:
select user_id, act_status, count(act_Status)
from your_table
where act_date in (
select max(act_date)
from your_table
group by Client_id
)
group by user_id, act_status
which when i add more transactions to the database, the output went wrong like this..
user_id | act_status | Count(act_status)
1 | dealed | 2
1 | hot | 1
22 | cold | 1
25 | hot | 1 ** (this one shouldn't show up)
25 | dealed | 1
26 | hot | 2 ** (should be only '1')
26 | dealed | 2

Answer 1

Tried and tested:
SELECT S.user_id, S.act_status, Count(S.act_status) AS CountOfact_status
FROM tbl_sActivity AS S INNER JOIN [SELECT A.client_id, Max(A.act_date) AS max_act
FROM tbl_sActivity AS A
GROUP BY A.client_id]. AS S2 ON (S.act_date = S2.max_act) AND (S.client_id = S2.client_id)
GROUP BY S.user_id, S.act_status
A similar request to an answer I gave yesterday, please try and read through my query as well as just using it, this will hopefully help you understand and is the only way to learn : )

Answer 2

Will it not be better if you use the primary key column: act_id, instead of using the date column.
Thus, getting the max primary key?
What I meant was, instead of this:
where act_date in (
select max(act_date)
from your_table
group by Client_id
)
something like this:
where act_id in (
select max(act_id)-- for each user, can't think of syntax now, but this could point you into the right direction
)
BugFinder's answer looks better.

Answer 3

select user_id, act_status, count(act_Status)
from your_table left join
(select act_date in (
select max(act_date)
from your_table
group by Client_id
) as t where t.client_id = your_table.client_id and t.act_date = your_table.act_date
group by user_id, act_status
Should give you what you wanted
Eg make a table of max dates by client, then link it back to your table using both fields to be precise.