我需要在报表中显示格式化的数据。我正在使用三个表，其中一个添加新用户，第二个添加新用户，最后一个添加新用户。因此，这是示例表和默认数据：
CREATE TABLE [dbo].[Entry]
(
[EmpId] [int] IDENTITY(1,1) NOT NULL,
[EmpName] [nvarchar](40) NULL,
[Address] [nvarchar](100) NULL,
[Email] [nvarchar](20) NULL,
[EntryDate] [datetime] NULL
)
INSERT [dbo].[Entry] ([EmpId], [EmpName], [Address], [Email], [EntryDate])
VALUES (1, N'John', N'On Earth', N'john@abc.com', CAST(0x0000A58000000000 AS DateTime)),
(2, N'Jack', N'On Earth', N'jack@abc.com', CAST(0x0000A5A800000000 AS DateTime)),
(3, N'Jessi', N'On Earth', N'jessi@abc.com', CAST(0x0000A5CF00000000 AS DateTime)),
(4, N'Jackson', N'On Earth', N'jackson@abc.com', CAST(0x0000A5E400000000 AS DateTime))
CREATE TABLE [dbo].[Payment]
(
[Id] [int] IDENTITY(1,1) NOT NULL,
[EmpId] [int] NULL,
[Payment] [float] NULL,
[PayDate] [datetime] NULL
)
INSERT [dbo].[Payment] ([Id], [EmpId], [Payment], [PayDate])
VALUES (1, 1, 2000, CAST(0x0000A61800000000 AS DateTime)), (2, 1, 2000, CAST(0x0000A63600000000 AS DateTime)),
(3, 1, 2000, CAST(0x0000A65500000000 AS DateTime)), (4, 1, 2000, CAST(0x0000A67400000000 AS DateTime)),
(5, 2, 4000, CAST(0x0000A5DB00000000 AS DateTime)), (6, 2, 4000, CAST(0x0000A5F900000000 AS DateTime)),
(7, 2, 4000, CAST(0x0000A61800000000 AS DateTime)), (8, 2, 4000, CAST(0x0000A63600000000 AS DateTime)),
(9, 2, 4000, CAST(0x0000A65500000000 AS DateTime)), (10, 2, 4000, CAST(0x0000A67400000000 AS DateTime)),
(11, 2, 4000, CAST(0x0000A69200000000 AS DateTime)), (12, 3, 6000, CAST(0x0000A65500000000 AS DateTime)),
(13, 3, 6000, CAST(0x0000A67400000000 AS DateTime)), (14, 4, 8000, CAST(0x0000A7FF00000000 AS DateTime)),
(15, 4, 8000, CAST(0x0000A98B00000000 AS DateTime))
CREATE TABLE [dbo].[Resign]
(
[Id] [int] IDENTITY(1,1) NOT NULL,
[EmpId] [int] NULL,
[ResignDate] [datetime] NULL,
[Reason] [nchar](10) NULL
)
INSERT [dbo].[Resign] ([Id], [EmpId], [ResignDate], [Reason])
VALUES (1, 1, CAST(0x0000A69B00000000 AS DateTime), N'Resigned '),
(2, 2, CAST(0x0000A6C400000000 AS DateTime), N'Resigned')
要求的输出如下，按年份：
Year 2015
Month - Jan - Feb - Mar - Apr - May - Jun - Jul - Aug - Sep - Oct - Nov - Dec
Opening - 10 - 10 - 10 - 12 - 12 - 12 - 12 - 12 - 12 - 12 - 8 - 8 //Total no. of employee
Add - 0 - 0 - 2 - 0 - 0 - 0 - 0 - 0 - 0 - 0 - 0 - 0 //Newly added
Left - 0 - 0 - 0 - 0 - 0 - 0 - 0 - 0 - 0 - 4 - 0 - 0 //Resigned
Closing - 10 - 10 - 12 - 12 - 12 - 12 - 12 - 12 - 12 - 8 - 8 - 8 //Closing Total
Year 2016
Month - Jan - Feb - Mar - Apr - May - Jun - Jul - Aug - Sep - Oct - Nov - Dec
Opening - 8 ----Goes on in the similar way
Add -----
Left ----
Closing ----
因此，很清楚应该如何在报告中显示数据。尽管我知道报告中按年份分组，并尝试类似地按月显示以下数据：
SELECT
DATEPART(YYYY, Q.PayDate) [YEAR],
COUNT(m.EmpId) Total,
SUM(CASE WHEN DATENAME(MONTH, m.EntryDate) = 'January' THEN 1 ELSE 0 END) [JAN],
SUM(CASE WHEN DATENAME(MONTH, m.EntryDate) = 'February' THEN 1 ELSE 0 END) [FEB]
FROM
Entry m
INNER JOIN
Payment q ON Q.EmpId = M.EmpId
LEFT JOIN
Resign OUT ON OUT.EmpId = m.EmpId
WHERE
m.EmpId NOT IN (SELECT DISTINCT p.EmpId FROM Resign p)
AND m.EmpId IN (SELECT DISTINCT k.EmpId FROM Payment k)
GROUP BY
DATEPART(YYYY, Q.PayDate), DATEPART(MM, m.EntryDate);
上面的查询返回如下内容：
YEAR Total JAN FEB
--------------------
2016 2 0 0
2017 1 0 0
2018 1 0 0
但是我不确定如何在与给定示例相同的查询中按行显示新添加的和辞职的用户？
NB ：我可以使用任何查询语言。因此，期望您的想法能够以适当的方式实施。
答案 0 :(得分：1)
Oracle解决方案。在第二行中放入报告年份。
mc每月统计新离职人员t是层次结构子查询，每月统计opening和closing的值。第一部分，锚点，对较早年份的表entry和resign中的条目进行计数，这是我们的出发点。对于该锚点，我们逐月附加在mc pivot和排序，仅在此处添加以根据您的请求显示数据。查询：
with
y as (select 2016 yr from dual),
mc as (
select mm, sum(a) a, sum(c) c
from ( select extract (month from entrydate) mm, 1 a, 0 c
from entry, y where extract(year from entrydate) = yr
union all
select extract (month from resigndate) mm, 0 a, 1 c
from resign, y where extract(year from resigndate) = yr)
group by mm ),
t(mm, o, a, r, c) as (
select 0, 0, 0, 0,
( (select count(1) from entry where extract(year from entrydate) < yr)
- (select count(1) from resign where extract(year from resigndate) < yr) )
from y
union all
select t.mm + 1, t.c, mc.a, mc.c, t.c + nvl(mc.a, 0) - nvl(mc.c, 0)
from t left join mc on mc.mm = t.mm + 1
where t.mm + 1 <= 12)
select *
from (
select * from t
unpivot (cnt for type in (o, a, r, c)))
pivot (sum(cnt) for mm in (1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12))
order by case type when 'O' then 1 when 'A' then 2
when 'R' then 3 when 'C' then 4
end
提供的数据和2016年的结果：
TYPE 1 2 3 4 5 6 7 8 9 10 11 12
---- ------ ------ ------ ------ ------ ------ ------ ------ ------ ------ ------ ------
O 0 1 2 3 4 4 4 4 4 4 3 2
A 1 1 1 1 0 0
R 0 0 0 0 1 1
C 1 2 3 4 4 4 4 4 4 3 2 2
主要是标准SQL，您可能需要将extract之类的更改功能更改为等效的SQLServer。希望这会有所帮助。
答案 1 :(得分：1)
如果您计划运行许多此类工具，则绝对应该将精力转移到诸如SSRS之类的报告工具上。它将简化很多次。话虽如此，这是一种产生您上面提到的结果的方法。
DECLARE @Entry TABLE
(
[EmpId] [int] NOT NULL,
[EmpName] [nvarchar](40) NULL,
[Address] [nvarchar](100) NULL,
[Email] [nvarchar](20) NULL,
[EntryDate] [datetime] NULL
)
INSERT @Entry ([EmpId], [EmpName], [Address], [Email], [EntryDate])
VALUES
(100, N'Early', N'On Earth', N'john@abc.com', '01/01/2015'),
(1, N'John', N'On Earth', N'john@abc.com', CAST(0x0000A58000000000 AS DateTime)),
(2, N'Jack', N'On Earth', N'jack@abc.com', CAST(0x0000A5A800000000 AS DateTime)),
(3, N'Jessi', N'On Earth', N'jessi@abc.com', CAST(0x0000A5CF00000000 AS DateTime)),
(4, N'Jackson', N'On Earth', N'jackson@abc.com', CAST(0x0000A5E400000000 AS DateTime))
DECLARE @Payment TABLE
(
[Id] [int] NOT NULL,
[EmpId] [int] NULL,
[Payment] [float] NULL,
[PayDate] [datetime] NULL
)
INSERT @Payment ([Id], [EmpId], [Payment], [PayDate])
VALUES (1, 1, 2000, CAST(0x0000A61800000000 AS DateTime)), (2, 1, 2000, CAST(0x0000A63600000000 AS DateTime)),
(3, 1, 2000, CAST(0x0000A65500000000 AS DateTime)), (4, 1, 2000, CAST(0x0000A67400000000 AS DateTime)),
(5, 2, 4000, CAST(0x0000A5DB00000000 AS DateTime)), (6, 2, 4000, CAST(0x0000A5F900000000 AS DateTime)),
(7, 2, 4000, CAST(0x0000A61800000000 AS DateTime)), (8, 2, 4000, CAST(0x0000A63600000000 AS DateTime)),
(9, 2, 4000, CAST(0x0000A65500000000 AS DateTime)), (10, 2, 4000, CAST(0x0000A67400000000 AS DateTime)),
(11, 2, 4000, CAST(0x0000A69200000000 AS DateTime)), (12, 3, 6000, CAST(0x0000A65500000000 AS DateTime)),
(13, 3, 6000, CAST(0x0000A67400000000 AS DateTime)), (14, 4, 8000, CAST(0x0000A7FF00000000 AS DateTime)),
(15, 4, 8000, CAST(0x0000A98B00000000 AS DateTime))
DECLARE @Resign TABLE
(
[Id] [int] NOT NULL,
[EmpId] [int] NULL,
[ResignDate] [datetime] NULL,
[Reason] [nchar](10) NULL
)
INSERT @Resign ([Id], [EmpId], [ResignDate], [Reason])
VALUES (1, 1, CAST(0x0000A69B00000000 AS DateTime), N'Resigned '),
(2, 2, CAST(0x0000A6C400000000 AS DateTime), N'Resigned')
DECLARE @StartDate DATETIME = '01/01/2015'
DECLARE @EndDate DATETIME = '12/01/2016'
;WITH Calendar as
(
SELECT CalendarDate = @StartDate, CalendarYear = DATEPART(YEAR, @StartDate), CalendarMonth = DATEPART(MONTH, @StartDate)
UNION ALL
SELECT CalendarDate = DATEADD(MONTH, 1, CalendarDate), CalendarYear = DATEPART(YEAR, CalendarDate), CalendarMonth = DATEPART(MONTH, CalendarDate)
FROM Calendar
WHERE DATEADD (MONTH, 1, CalendarDate) <= @EndDate
)
,Employees AS
(
SELECT
E.EmpID,
StartDate = DATEADD(MONTH, DATEDIFF(MONTH, 0, EntryDate), 0),
EndDate = DATEADD(MONTH, DATEDIFF(MONTH, 0, ResignDate), 0)
FROM
@Entry E
LEFT OUTER JOIN @Resign R ON R.EmpId = E.EmpId
)
,NormalizedAndUnpivoted AS
(
SELECT
*,
CalendarMonth = DATEPART(MONTH,CalendarDate),
CalendarYear = DATEPART(YEAR,CalendarDate),
GroupField = CAST(DATEPART(MONTH,CalendarDate) AS NVARCHAR(50))+'_'+CAST(DATEPART(YEAR,CalendarDate) AS NVARCHAR(50)),
SortOrder = CASE
WHEN Property='Opening' THEN 1
WHEN Property='Addition' THEN 2
WHEN Property='Subtraction' THEN 3
WHEN Property='Total' THEN 4
END
FROM
(
SELECT
C.CalendarDate,
Opening = COUNT(DISTINCT EmpExisting.EmpID) ,
Addition = COUNT(DISTINCT EmpStarted.EmpID),
Subtraction = COUNT(DISTINCT EmpEnd.EmpID),
Total = COUNT(DISTINCT EmpExisting.EmpID) + COUNT(DISTINCT EmpStarted.EmpID) - COUNT(DISTINCT EmpEnd.EmpID)
FROM
Calendar C
LEFT OUTER JOIN Employees EmpExisting ON EmpExisting.StartDate < C.CalendarDate AND (EmpExisting.EndDate IS NULL OR EmpExisting.EndDate >= C.CalendarDate)
LEFT OUTER JOIN Employees EmpStarted ON EmpStarted.StartDate = C.CalendarDate
LEFT OUTER JOIN Employees EmpEnd ON EmpEnd.EndDate = C.CalendarDate
GROUP BY
C.CalendarDate
)AS X
UNPIVOT(
PivotValue FOR Property IN (Opening, Addition, Subtraction, Total)
) AS U
)
,Pivoted AS
(
SELECT
CalendarYear,
Property,
SortOrder,
Janurary=SUM(CASE WHEN CalendarMonth=1 THEN PivotValue ELSE NULL END),
Feburary=SUM(CASE WHEN CalendarMonth=2 THEN PivotValue ELSE NULL END),
March=SUM(CASE WHEN CalendarMonth=3 THEN PivotValue ELSE NULL END),
April=SUM(CASE WHEN CalendarMonth=4 THEN PivotValue ELSE NULL END),
May=SUM(CASE WHEN CalendarMonth=5 THEN PivotValue ELSE NULL END),
June=SUM(CASE WHEN CalendarMonth=6 THEN PivotValue ELSE NULL END),
July=SUM(CASE WHEN CalendarMonth=7 THEN PivotValue ELSE NULL END),
August=SUM(CASE WHEN CalendarMonth=8 THEN PivotValue ELSE NULL END),
September=SUM(CASE WHEN CalendarMonth=9 THEN PivotValue ELSE NULL END),
October=SUM(CASE WHEN CalendarMonth=10 THEN PivotValue ELSE NULL END),
November=SUM(CASE WHEN CalendarMonth=11 THEN PivotValue ELSE NULL END),
December=SUM(CASE WHEN CalendarMonth=12 THEN PivotValue ELSE NULL END)
FROM
NormalizedAndUnpivoted
GROUP BY
CalendarYear,
Property,
SortOrder
)
SELECT * FROM Pivoted
ORDER BY
CalendarYear,
SortOrder
OPTION (MAXRECURSION 1000)
结果：
CalendarYear Property SortOrder Janurary Feburary March April May June July August September October November December
------------ ------------------------------ ----------- ----------- ----------- ----------- ----------- ----------- ----------- ----------- ----------- ----------- ----------- ----------- -----------
2015 Opening 1 0 1 1 1 1 1 1 1 1 1 1 1
2015 Addition 2 1 0 0 0 0 0 0 0 0 0 0 0
2015 Subtraction 3 0 0 0 0 0 0 0 0 0 0 0 0
2015 Total 4 1 1 1 1 1 1 1 1 1 1 1 1
2016 Opening 1 1 2 3 4 5 5 5 5 5 5 4 3
2016 Addition 2 1 1 1 1 0 0 0 0 0 0 0 0
2016 Subtraction 3 0 0 0 0 0 0 0 0 0 1 1 0
2016 Total 4 2 3 4 5 5 5 5 5 5 4 3 3