SQL need to compare rows count values - sql

Question

I have a query that returns the ID, Name and count of the number of times an ID has been entered to the table.
SELECT
ID,
NAME,
COUNT(*) count
FROM
TABLE
GROUP BY
NAME, ID, CASE_DETAIL_ID
HAVING
COUNT(*) > 1;
This returns the following data:
ID
NAME
COUNT
123
HAT
10
123
UMBRELLA
10
123
TOWEL
10
123
WATER
8
555
HAT
3
555
UMBRELLA
10
555
TOWEL
10
555
WATER
10
322
UMBRELLA
5
322
TOWEL
20
322
WATER
20
I want to be able to query the row with a count of less than what the other rows with the same ID have. How can I do this? So that the end result is:
ID
NAME
COUNT
FULL COUNT
123
WATER
8
10
555
HAT
3
10
322
UMBRELLA
5
20
There are multiple IDs that we store and I only want the rows/names that have a count less than the rows with the same IDs have.
I have also tried -
WITH x AS
(SELECT ID, NAME, COUNT(*) count
FROM FRT.CASE_DETAIL_HISTORY
GROUP BY
NAME,
ID,
CASE_DETAIL_ID)
SELECT x.ID, t.NAME, X.COUNT, MIN(x.count)
FROM x
JOIN FRT.CASE_DETAIL_HISTORY t
on t.ID= x.ID
GROUP BY x.ID, t.ID, X.COUNT
However, this doesnt give me what I am looking for. I only want rows returned if the name's count doesnt match the 'mode' count of the ID.
I also have tried the below but keep facing errors:
WITH COUNT_OF_ROWS AS
(SELECT ID, NAME, COUNT(*) count
FROM TABLE
GROUP BY NAME, ID, CASE_DETAIL_ID
HAVING COUNT(*) >= 1),
MINIMUM AS
(SELECT COUNT_OF_ROWS.ID, COUNT_OF_ROWS.NAME,
MIN(COUNT_OF_ROWS.COUNT) MINI
FROM COUNT_OF_ROWS
JOIN TABLE CD on CD.ID = COUNT_OF_ROWS.ID
GROUP BY COUNT_OF_ROWS.ID, COUNT_OF_ROWS.NAME
)
select distinct COUNT_OF_ROWS.*, MINIMUM.MINI
from minimum, count_of_rows
where minimum.mini != count_of_rows.count;

Answer 1

Some sample data would help but you can use a CTE, and select the lowest using min() something like this:
WITH x AS(
SELECT t.id, t.nametext, COUNT(*) as count
FROM table t
GROUP BY id, t.nametext, CASE_DETAIL_ID
), y as(
SELECT x.id, MIN(x.[COUNT]) as mincount
FROM x
GROUP BY x.id
)
select y.id, x.nametext, y.mincount
from y
join x
on x.[COUNT] = y.mincount
and x.id = y.id
Or it can be done using top 1 and order by like this:
SELECT TOP 1 id, name, COUNT(*) as count
FROM TABLE
WHERE ID = 123
GROUP BY NAME, ID, CASE_DETAIL_ID
ORDER BY count DESC
But bare in mind that as this would select only the first row, this would only work with the where clause because it would always only return 1 row.
While if you use the CTE option it would work also if you want per id, without where id = 123.