对每个ID的每一行计数一次，并获得计数SQL Select查询

Question

So, I have a series of ids in a table with attributes cid, wid, did, where entries in this table can be null, but if they are not null then the entries along a row will be equal. This corresponds to the ids of members of a film crew, writers and directors ids.

What I'm looking to do is to choose those ids who have worked on 3 or more films. Thinking about it as a for loop, I'm imagining I have a dictionary with the key as the id, and I want to iterate through each row in the array, adding one to whatever id it is in the row's dictionary, and then I can just use a HAVING clause to select out those with COUNT >= 3.

My major problem is that I don't want to count the id more than once even if it appears multiple times in a row.

Here is a sample of the data:

   cid    |   did    |   wid
----------+----------+----------
 00000027 |          |
 00000209 |          |
 00000205 |          |
 00000206 |          |
 00000207 |          |
 00000208 |          |
 00001140 |          |
 00000306 |          |
 00000307 |          |
 00000325 |          |
 00000349 |          |
 00001077 |          |
 00001078 |          |
 00001079 | 00001079 | 00001079
 00001079 | 00001079 | 00001079
 00001082 |          |
 00001083 |          |
 00001084 |          |
 00001085 |          |
 00001091 |          |
 00001092 |          |
 00001093 |          |
 00001094 |          |
 00001095 |          |
 00001096 | 00001096 |
 00001101 |          |
 00001102 |          |
 00001104 | 00001104 | 00001104
 00001104 | 00001104 | 00001104
 00001104 | 00001104 | 00001104
 00001314 |          |
 00001315 |          |
 00001316 |          |
 00001321 |          |
 00001322 |          |
 00001323 |          |
 00001328 |          |
 00001329 |          |
 00001330 |          |
 00001335 |          |
 00001336 |          |
 00001337 |          |
 00001345 |          |
 00001345 |          |
          |          | 00001344
          |          | 00000461
          |          | 00001344
          |          | 00001332

Answer 1

SELECT cid, count(*) "nrOfFilms"
  FROM people
 GROUP BY cid
HAVING nrOfFilms >= 3

Answer 2

You may try this, first you need to combine all of the columns into one on basis of any non value, as per your statement. After that you can easily apply group by and having to get your desired result.

; with cte as (
select case when isnull(cid,0)=0 
            then case when isnull(did,0)=0 
                 then isnull(wid,0) 
                 else isnull(did,0) end
            else isnull(cid,0) end as Uid  from table
)
select distinct uid, count(1) as CountId from cte 
group by uid having count(uid)>2