我想将 ROLLUP 与 PIVOT 结合起来 - 这是一个选择吗? [英] I would like to combine ROLLUP with PIVOT - is that an option?

问题描述

我一直在用

SELECT
    Author,
    ISNULL(MAX(CASE Status WHEN 'Duplicate' THEN NumDocs END),'') AS Duplicate,
    ISNULL(MAX(CASE Status WHEN 'Failure' THEN NumDocs END),'') AS Failure,
    ISNULL(MAX(CASE Status WHEN 'Rejected' THEN NumDocs END),'') AS Rejected,
    ISNULL(MAX(CASE Status WHEN 'Success' THEN NumDocs END),'') AS Success,
    ISNULL(MAX(CASE Status WHEN 'TOTAL' THEN NumDocs END),'') AS TOTAL
FROM    
    (SELECT
        CASE WHEN (GROUPING(Author)=1) THEN 'ALL'
            ELSE ISNULL(Author,'UNKNOWN') END AS Author,
        CASE WHEN (GROUPING(Status )=1) THEN 'TOTAL'
            ELSE ISNULL(Status ,'UNKNOWN') END AS [Status],
        COUNT(Status) AS NumDocs
    FROM 
        tbl_Document D
    LEFT JOIN
        tbl_Status S
            ON
                D.status_id = S.status_id   
    GROUP BY
        Author,
        Status
    WITH ROLLUP) BASE
GROUP BY 
    Author

转换:

[Author]  [Status]
Alan      SUCCESS
Bob       FAILURE
Bob       SUCCESS
Charles   SUCCESS
Dave      FAILURE
Dave      DUPLICATE

到:

[Author] [SUCCESS] [FAILURE] [DUPLICATE] [TOTALS]
Alan        1         0           0         1
Bob         1         1           0         2
Charles     1         0           0         1
Dave        0         1           1         2
TOTAL       3         2           1         6

我可以使用 PIVOT 语句接近此输出，但我不确定如何获取 TOTAL 行/列?

I can get close to this output using a PIVOT statement, but I'm not sure how to get the TOTAL row/column?

SELECT
  * 
FROM 
  (SELECT Author, status_id FROM tbl_Document) d
PIVOT
  (COUNT(status_id) FOR status_id IN ([1],[3],[5],[6])) p

给予:

[Author] [SUCCESS] [FAILURE] [DUPLICATE] 
Alan        1         0           0      
Bob         1         1           0      
Charles     1         0           0      
Dave        0         1           1     

我猜我需要将 ROLLUP 放入某个子查询中...?

I'm guessing I need to put the ROLLUP into a subquery somewhere...?

推荐答案

你没有发布表架构，所以我尝试推断它.我从您提供的输入开始(请参阅最里面的 SELECT 中的注释)，因此您应该能够根据您的实际模式进行调整.我在没有任何文档的情况下添加了一个额外的作者，因为我认为您希望在最终报告输出中看到这些作者.排除这些作者很简单.

You didn't post the table schema, so I tried to infer it. I started with the input you gave (see the comment in the innermost SELECT), so you should be able to adapt this to your actual schema. I included an extra author without any documents, because I figured you'd want to see those in the final report output. It's trivial to exclude those authors.

DECLARE @Status table
(
    Id int NOT NULL,
    Status nvarchar(50) NOT NULL
)

DECLARE @Authors table
(
    Id int NOT NULL,
    Name nvarchar(50) NOT NULL
)

DECLARE @Documents table
(
    Id int NOT NULL,
    AuthorId int NOT NULL,
    StatusId int NOT NULL
)

INSERT INTO @Status VALUES (1, 'Duplicate')
INSERT INTO @Status VALUES (2, 'Failure')
INSERT INTO @Status VALUES (3, 'Rejected')
INSERT INTO @Status VALUES (4, 'Success')

INSERT INTO @Authors VALUES (1, 'Alan')
INSERT INTO @Authors VALUES (2, 'Bob')
INSERT INTO @Authors VALUES (3, 'Charles')
INSERT INTO @Authors VALUES (4, 'Dave')
INSERT INTO @Authors VALUES (5, 'Tom') -- Test for authors without documents

INSERT INTO @Documents VALUES (1, 1, 4)
INSERT INTO @Documents VALUES (2, 2, 2)
INSERT INTO @Documents VALUES (3, 2, 4)
INSERT INTO @Documents VALUES (4, 3, 4)
INSERT INTO @Documents VALUES (5, 4, 2)
INSERT INTO @Documents VALUES (6, 4, 1)

SELECT
    (CASE WHEN GROUPING(Name) = 1 THEN 'Total' ELSE Name END) AS Author,
    SUM(Duplicate) AS Duplicate,
    SUM(Failure) AS Failure,
    SUM(Rejected) AS Rejected,
    SUM(Success) AS Success,
    SUM(Duplicate + Failure + Rejected + Success) AS Total
    FROM
    (
        SELECT
            Name,
            (CASE WHEN Status = 'Duplicate' THEN 1 ELSE 0 END) AS Duplicate,
            (CASE WHEN Status = 'Failure' THEN 1 ELSE 0 END) AS Failure,
            (CASE WHEN Status = 'Rejected' THEN 1 ELSE 0 END) AS Rejected,
            (CASE WHEN Status = 'Success' THEN 1 ELSE 0 END) AS Success
            FROM
            (
                -- Original input
                SELECT
                    a.Name,
                    s.Status
                    FROM @Authors a
                    LEFT OUTER JOIN @Documents d ON d.AuthorId = a.Id
                    LEFT OUTER JOIN @Status s ON d.StatusId = s.Id
            ) i
    ) j
    GROUP BY Name WITH ROLLUP

输出:

Author   Duplicate  Failure  Rejected  Success  Total
Alan     0          0        0         1        1
Bob      0          1        0         1        2
Charles  0          0        0         1        1
Dave     1          1        0         0        2
Tom      0          0        0         0        0
Total    1          2        0         3        6

这篇关于我想将 ROLLUP 与 PIVOT 结合起来 - 这是一个选择吗?的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！