SQL DATEDIFF返回

Question

I don't want to show my whole query as it is very specific but I will try to explain briefly.

The following query works perfectly and I got 6000 records as a result.

  SELECT
      DISTINCT ID,
      NAME,
      CASE WHEN A.ID IS NULL THEN 'NOT EX.'
       ELSE A.Whatever
      END AS A_Whatever,
      D.Z1 AS A.P 
      --AND SO ON......          

  FROM A 

      INNER JOIN A ON A.ID= B.ID AND A.Nb= B.Nb
      LEFT JOIN T AS T2_ID ON T2_D.Z= A.Z
      LEFT JOIN L1 ON A.NR = L1.NR AND A.S = L1.S
      LEFT JOIN LF ON LF.NR = L1.LNR
      --AND SO ON.......

However, when i add a DATEDIFF calculation I get only 100 Rows in the answer:

  SELECT DISTINCT 
      ID,
      NAME,
      CASE WHEN A.ID IS NULL THEN 'NOT EX.'
        ELSE A.Whatever
      END AS A_Whatever,
      D.Z1 AS A.P,
      DATEDIFF(dd,A.ADATE,A.BDATE)

FROM A 
  INNER JOIN A ON A.ID= B.ID AND A.Nb= B.Nb
  LEFT JOIN T AS T2_ID ON T2_D.Z= A.Z
  LEFT JOIN L1 ON A.NR = L1.NR AND A.S = L1.S
  LEFT JOIN LF ON LF.NR = L1.LNR
  --AND SO ON.......

I am expecting 6000 rows with a correct query using DATEDIFF in line with what the following query returns:

  SELECT 
      DISTINCT *,
      DATEDIFF(dd,A.ADATE,A.BDATE)
  FROM A 
      INNER JOIN A ON A.ID= B.ID AND A.Nb= B.Nb
      LEFT JOIN T AS T2_ID ON T2_D.Z= A.Z
      LEFT JOIN L1 ON A.NR = L1.NR AND A.S = L1.S
      LEFT JOIN LF ON LF.NR = L1.LNR
      --AND SO ON....... 

But I do not need all of them I need just the selected ones and the DATEDIFF but combining the queries above did not work for some reason that i do not know. Can anyone see why i am not getting the expected row count in my second query?

Answer 1

You can just wrap your query in a CTE and perform the DATEDIFF on the result set returned by the CTE :

WITH DISTINCT_CTE AS (
   SELECT
         DISTINCT ID,
         NAME,
         CASE WHEN A.ID IS NULL THEN 'NOT EX.'
              ELSE A.Whatever
         END AS A_Whatever,
         D.Z1 AS A.P 
         --AND SO ON......              
   FROM A 

        INNER JOIN A ON A.ID= B.ID AND A.Nb= B.Nb
        LEFT JOIN T AS T2_ID ON T2_D.Z= A.Z
        LEFT JOIN L1 ON A.NR = L1.NR AND A.S = L1.S
        LEFT JOIN LF ON LF.NR = L1.LNR
           --AND SO ON.......
)
SELECT *, DATEDIFF(dd, ADATE, BDATE)
FROM DISTINCT_CTE

Answer 2

You could try a subquery... first subquery the entire thing.

In the receiving query do another subquery for the DATEDIFF. For this subquery you need the primary key to get back to the correct row for the dates, which if I'm interpreting correctly is A.ID .

SELECT dT.ID, dT.Name, dT.A_Whatever           
      ,(SELECT DATEDIFF(dd, A2.ADATE, A2.BDATE)
          FROM A AS A2
         WHERE dT.ID = A2.ID --the primary key
       ) AS [DateDiff]
        --AND SO ON........
  FROM ( 
         SELECT DISTINCT ID,
                NAME,
                CASE WHEN A.ID IS NULL THEN 'NOT EX.'
                                       ELSE A.Whatever
                END AS A_Whatever,
                D.Z1 AS A.P 
               --AND SO ON......          

         FROM A 

             INNER JOIN A ON A.ID= B.ID AND A.Nb= B.Nb
             LEFT JOIN T AS T2_ID ON T2_D.Z= A.Z
             LEFT JOIN L1 ON A.NR = L1.NR AND A.S = L1.S
             LEFT JOIN LF ON LF.NR = L1.LNR
             --AND SO ON.......
      ) AS dT

Answer 3

You can try the following query :

SELECT T1.*, T2.DateDiff
FROM ( 
    SELECT DISTINCT ID,
            NAME,
            CASE WHEN A.ID IS NULL THEN 'NOT EX.'
                                   ELSE A.Whatever
            END AS A_Whatever,
            D.Z1 AS A.P 
           --AND SO ON......          
    FROM A 
         INNER JOIN B ON A.ID= B.ID AND A.Nb= B.Nb
         LEFT JOIN T AS T2_ID ON T2_D.Z= A.Z
         LEFT JOIN L1 ON A.NR = L1.NR AND A.S = L1.S
         LEFT JOIN LF ON LF.NR = L1.LNR
         --AND SO ON.......
  ) AS T1

JOIN (SELECT ID, DATEDIFF(dd, ADATE, BDATE) as DateDiff from A as A2) AS T2
ON T1.ID= T2.ID

Answer 4

As the following attempts to display, if you are using select distinct AND removing date columns from view when introducing datediff() into the select clause that could be the cause of the change in rows returned. Note in query 1 that as long as adate or bdate are displayed 5 rows would be returned, but without them (query 2) you just get one row. Alternatively, if you removed distinct,from query 2 you would get all 5 rows but just the one column (this isn't shown below).

try it out at SQL Fiddle

MS SQL Server 2014 Schema Setup :

CREATE TABLE Sample
    ([ADATE] datetime, [BDATE] datetime)
;

INSERT INTO Sample
    ([ADATE], [BDATE])
VALUES
    ('2017-10-01 00:00:00', '2017-10-06 00:00:00'),
    ('2017-10-02 00:00:00', '2017-10-07 00:00:00'),
    ('2017-10-03 00:00:00', '2017-10-08 00:00:00'),
    ('2017-10-04 00:00:00', '2017-10-09 00:00:00'),
    ('2017-10-05 00:00:00', '2017-10-10 00:00:00')
;

Query 1 :

select distinct 'q1' qry, adate, bdate, datediff(day,adate,bdate) days_diff
from sample
order by adate

Results :

| qry |                adate |                bdate | days_diff |
|-----|----------------------|----------------------|-----------|
|  q1 | 2017-10-01T00:00:00Z | 2017-10-06T00:00:00Z |         5 |
|  q1 | 2017-10-02T00:00:00Z | 2017-10-07T00:00:00Z |         5 |
|  q1 | 2017-10-03T00:00:00Z | 2017-10-08T00:00:00Z |         5 |
|  q1 | 2017-10-04T00:00:00Z | 2017-10-09T00:00:00Z |         5 |
|  q1 | 2017-10-05T00:00:00Z | 2017-10-10T00:00:00Z |         5 |

Query 2 :

select distinct 'q2' qry, datediff(day,adate,bdate) days_diff
from sample

Results :

| qry | days_diff |
|-----|-----------|
|  q2 |         5 |