Welcome to OGeek Q&A Community for programmer and developer-Open, Learning and Share
Welcome To Ask or Share your Answers For Others

Categories

0 votes
424 views
in Technique[技术] by (71.8m points)

sql server 2008 - SQL - Pivot table and group by not working

I have a table as follows:

Product     #users  Date            Count   Type
prodA       29      2013-02-27      113     Document
prodA       31      2013-03-02      152     Extraction
prodB       17      2013-02-26      40      Document
prodB       28      2013-03-02      73      Extraction

I need to use a pivot table on the [Type]/Count column and obtain a table as follows:

Product     #users  Date            Document Extraction
prodA       60      2013-03-02      113      152
prodB       45      2013-03-02      40       73

where the #user column is the sum group by product, and the Date is the max date group by product.

This is what I got so far:

SELECT Product, 
       sum(users),
       max([Date]),  
       [Document],[Extraction] FROM Table

     PIVOT 
        ( sum([Count]) FOR [Type] IN ( Document , Extraction)) AS [QUANTITY]

     GROUP BY activity, document, extraction

but my final results give me something like this instead:

Product     #users Date      Document Extraction
prodA       31     2013-03-02 NULL     152
prodA       29     2013-02-27 113      NULL
prodB       28     2013-03-02 NULL     73
prodB       17     2013-02-26 40       NULL

It is not grouping by the Product!

Any ideas?

EDIT:

So far, I have

WITH Pivoted
AS
(
  SELECT *
  FROM table1
  PIVOT 
  ( 
    sum([Count]) FOR [Type] IN ( Document , 
                                Extraction)
  ) AS p
) 
SELECT 
  product,
  SUM(Users) AS TotalUsers,
  MAX(DAte) AS LatestDate,
  MAX(Document) AS Document,
  MAX(Extraction) AS Extraction
FROM Pivoted
GROUP BY Product;

but my table1 above is actually made of codes below:

WITH a
AS(
SELECT    activity, 
                username, 
                [Last Accessed] = max(DATEADD(dd, DATEDIFF(d, 0, ActDateTime), 0)), --[#Users] = count(distinct username), 
                CASE WHEN COUNT(activity)IS NOT NULL THEN 1 ELSE 0 END AS Count,
                CASE WHEN pageURL LIKE '/Document%'
                OR pageURL LIKE '/Database%' THEN 'Document' ELSE 'Extraction' END AS [Type] --into #temp
                from activitylog
                where pageURL not like '%home%' AND pageURL not like '/Default%'
                --AND ActDateTime >= @StartDate AND ActDateTime <= @EndDate
                group by activity, 
                username, 
                --department,
                DATEADD(dd, DATEDIFF(d, 0, ActDateTime), 0), 
                CASE WHEN pageURL LIKE '/Document%' OR pageURL LIKE '/Database%' THEN 'Document' ELSE 'Extraction' END 
                --order by activity--, username, department,DATEADD(dd, DATEDIFF(d, 0, ActDateTime), 0)
    )

    ,b as
    (select activity, count(distinct username) as [Users] ,
        --department ,
        max([Last Accessed]) as [Last Accessed1],count([count])as [Count],[Type] from a --into #temp1 from #temp
        group by activity, 
        --department,
        [Type]

        )  

        select * from b order by activity;

so my question is, HOW do i put the chunk of code above that makes Table1 within the first WITH AS?

Thanks

See Question&Answers more detail:os

与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…
Welcome To Ask or Share your Answers For Others

1 Reply

0 votes
by (71.8m points)

You can't GROUP BY activity, document, extraction within the PIVOT table operator, the PIVOT operator infers the grouped columns automatically. But you can write it this way:

WITH Pivoted
AS
(
  SELECT *
  FROM table1
  PIVOT 
  ( 
    sum([Count]) FOR [Type] IN ( Document , 
                                Extraction)
  ) AS p
) 
SELECT 
  product,
  SUM(Users) AS TotalUsers,
  MAX(DAte) AS LatestDate,
  MAX(Document) AS Document,
  MAX(Extraction) AS Extraction
FROM Pivoted
GROUP BY Product;

SQL Fiddle Demo

This will give you:

| PRODUCT | TOTALUSERS |                   LATESTDATE | DOCUMENT | EXTRACTION |
-------------------------------------------------------------------------------
|   prodA |         60 | March, 02 2013 02:00:00+0000 |      113 |        152 |
|   prodB |         45 | March, 02 2013 02:00:00+0000 |       40 |         73 |

Update 1

WITH a
AS(
  SELECT    
    activity, 
    username, 
    [Last Accessed] = max(DATEADD(dd, 
                                  DATEDIFF(d, 0, ActDateTime), 
                                  0)), 
    --[#Users] = count(distinct username), 
    CASE 
      WHEN COUNT(activity) IS NOT NULL THEN 1 
      ELSE 0 
    END AS Count,
    CASE 
      WHEN pageURL LIKE '/Document%'
        OR pageURL LIKE '/Database%' THEN 'Document'
      ELSE 'Extraction' 
    END AS [Type] 
  from activitylog
  where pageURL not like '%home%' 
    AND pageURL not like '/Default%'
  group by activity, 
           username, 
           ...
), Pivoted
AS
(
  SELECT *
  FROM a
  PIVOT 
  ( 
    sum([Count]) FOR [Type] IN ( Document , 
                                 Extraction)
   ) AS p
) 
SELECT 
  product,
  SUM(Users) AS TotalUsers,
  MAX(DAte) AS LatestDate,
  MAX(Document) AS Document,
  MAX(Extraction) AS Extraction
FROM Pivoted
GROUP BY Product; 

与恶龙缠斗过久,自身亦成为恶龙;凝视深渊过久,深渊将回以凝视…
OGeek|极客中国-欢迎来到极客的世界,一个免费开放的程序员编程交流平台!开放,进步,分享!让技术改变生活,让极客改变未来! Welcome to OGeek Q&A Community for programmer and developer-Open, Learning and Share
Click Here to Ask a Question

...