Wie in mehreren Spalten zählen, eine Variable

Question

ich auf einem Crystal Report arbeite und kann nicht einen Weg finden, dies in Crystal oder MSSQL

ich eine Tabelle mit Benutzernamen zu tun haben, die aufgezeichnet werden, wenn eine Aktivität in abgeschlossen ist eine Web wie app unter

UID Approved Denied Moved Pended 
1 Tom   null null Bill 
2 null  null Bill null 
3 Bill  null Tom  null 
4 null  Tom  null null 
5 Tom   null Bill Bill 
6 null  Bill Bill null 
7 Tom   null null Bill 

ich brauche einen Produktivitäts Bericht wie

 Approved Denied Moved Pended 
Tom  3   1  1  0 
Bill 1   1  3  3 

ich bin sicher nicht, wie die Zählungen der Gruppe durch die Namen unten zu erzeugen, die in den Daten?

Answer 1

SQL Fiddle Demo

Zuerst sollten Sie alle Benutzer in eine temporäre Tabelle dann die Gruppe alle Aggregationen in einer Tabelle

--Get all users 

SELECT * 
INTO #Tbl_User 
FROM (

SELECT DISTINCT Approved as [User] 
FROM MY_TABLE 

UNION 

SELECT DISTINCT Denied as [User] 
FROM MY_TABLE 

UNION 

SELECT DISTINCT Moved as [User] 
FROM MY_TABLE 

UNION 

SELECT DISTINCT Pended as [User] 
FROm MY_TABLE) AS T1 WHERE T1.User IS NOT NULL 

--Group results 

SELECT T1.[User], ISNULL(T2.Approved,0) Approved, ISNULL(T3.Denied,0) Denied, ISNULL(T4.Moved,0) Moved, ISNULL(T5.Pended,0) Pended 
FROM #Tbl_User AS T1 INNER JOIN (
SELECT Approved as [User],Count(*) as Approved  
FROM MY_TABLE 
GROUP BY Approved) AS T2 ON T1.[User] =T2.[User] 
LEFT JOIN (
SELECT Denied as [User],Count(*) as Denied  
FROM MY_TABLE 
GROUP BY Denied) AS T3 ON T1.[User] =T3.[User] 
LEFT JOIN (
SELECT Moved as [User],Count(*) as Moved  
FROM MY_TABLE 
GROUP BY Moved) AS T4 ON T1.[User] =T4.[User] 
LEFT JOIN (
SELECT Pended as [User],Count(*) as Pended  
FROM MY_TABLE 
GROUP BY Pended) AS T5 ON T1.[User] =T5.[User] 

Answer 2

Wenn es nicht massiv ist, können Sie CTE gekettet verwenden:

DECLARE @activity AS TABLE 
    (
     [uid]  INT 
     , [approved] NVARCHAR(25) 
     , [denied] NVARCHAR(25) 
     , [moved] NVARCHAR(25) 
     , [pending] NVARCHAR(25) 
); 

INSERT INTO @activity 
      ([uid] 
      , [approved] 
      , [denied] 
      , [moved] 
      , [pending]) 
VALUES  (1,N'Curly',NULL,NULL,NULL), 
      (2,NULL,NULL,N'Curly',NULL), 
      (3,N'Curly',NULL,NULL,N'Mo'), 
      (4,NULL,NULL,N'Joe',NULL), 
      (5,N'Joe',NULL,N'Joe',NULL), 
      (6,N'Joe',N'Shemp',N'Joe',NULL); 

WITH [approved_builder] 
    AS (SELECT [approved] AS [name] 
       , count(*) AS [approved] 
     FROM @activity 
     WHERE [approved] IS NOT NULL 
     GROUP BY [approved]) 
    , [moved_builder] 
    AS (SELECT [moved] AS [name] 
       , count(*) AS [moved] 
     FROM @activity 
     WHERE [moved] IS NOT NULL 
     GROUP BY [moved]) 
    , [denied_builder] 
    AS (SELECT [denied] AS [name] 
       , count(*) AS [denied] 
     FROM @activity 
     WHERE [denied] IS NOT NULL 
     GROUP BY [denied]) 
    , [pending_builder] 
    AS (SELECT [pending] AS [name] 
       , count(*) AS [pending] 
     FROM @activity 
     WHERE [pending] IS NOT NULL 
     GROUP BY [pending]) 
    , [aggregator] 
    AS (SELECT [name] 
       , [approved] AS [approved] 
       , 0   AS [denied] 
       , 0   AS [moved] 
       , 0   AS [pending] 
     FROM [approved_builder] 
     UNION ALL 
     SELECT [name] 
       , 0  AS [approved] 
       , 0  AS [denied] 
       , [moved] AS [moved] 
       , 0  AS [pending] 
     FROM [moved_builder] 
     UNION ALL 
     SELECT [name] 
       , 0  AS [approved] 
       , [denied] AS [denied] 
       , 0  AS [moved] 
       , 0  AS [pending] 
     FROM [denied_builder] 
     UNION ALL 
     SELECT [name] 
       , 0   AS [approved] 
       , 0   AS [denied] 
       , 0   AS [moved] 
       , [pending] AS [pending] 
     FROM [pending_builder]) 
SELECT [name]   AS [name] 
     , sum([approved]) AS [approved] 
     , sum([denied]) AS [denied] 
     , sum([moved]) AS [moved] 
     , sum([pending]) AS [pending] 
FROM [aggregator] 
GROUP BY [name]; 

Answer 3

Ich denke, diese einfache Abfrage wird den Trick machen, um das gewünschte Ergebnis zu erhalten.

SELECT Approved as Names, count(Approved) as Approved, 
    (SELECT count(Denied) 
    FROM tableName AS j2 
    WHERE j1.Approved = j2.Denied) AS Denied, 
    (SELECT count(Moved) 
    FROM tableName AS j2 
    WHERE j1.Approved = j2.Moved) AS Moved, 
    (SELECT count(Pended) 
    FROM tableName AS j2 
    WHERE j1.Approved = j2.Pended) AS Pended 
FROM tableName AS j1 
GROUP BY Approved 

Answer 4

Sie können auch Dreh- und Entpivotisierung Syntax wie unten verwenden: See working demo

select 
Name, 
[approved]=ISNULL([approved],0), 
[denied]=ISNULL([denied],0), 
[moved]=ISNULL([moved],0), 
[pending]=ISNULL([pending],0) 
from 
(
    select 
     Name,activity,count(1) c 
    from 
    (select * from useractivity)s 
    unpivot (Name for activity in ([approved],[denied],[moved],[pending]))up 
    group by Name,activity 
) s 
pivot (sum(c) for activity in ([approved],[denied],[moved],[pending]))p 

Answer 5

Ja, es hat nicht verschwenkten und dann zu verschwenken. Hier ist, wie wir dies in Oracle SQL tun können.

SELECT * 
FROM 
    (SELECT name, 
    status 
    FROM report UNPIVOT(name FOR status IN (APPROVED AS 'APPROVED',DENIED AS 'DENIED', MOVED AS 'MOVED', PENDED AS 'PENDED')) 
) PIVOT (COUNT(status) FOR status IN ('APPROVED','DENIED','MOVED','PENDED')) 
ORDER BY name DESC;