我的SQL Server上有一个存储过程，它整合了一系列字段，以便在SSRS for Report Builder中使用。该过程由FileId提供，然后运行其逻辑。它按预期工作，直到文件无法找到或引用Solicitor和Arresting Officer字段。
即使fileId没有关联的逮捕官或律师，我也需要这个来返回结果。我确信这很简单。基本上，如果CTE从查询中返回任何内容，我仍然需要一个默认行。
ALTER PROCEDURE [dbo].[GetRollCallData]
@Ids VARCHAR(255),
@LexiconId INT,
@UUID UNIQUEIDENTIFIER,
@ReadOnly INT
AS
DECLARE @TableCode INT
SET @TableCode = 58
IF @Ids <> ''
BEGIN
EXEC InsertInSelectionCache @Ids, @UUID, @TableCode, 0
IF @ReadOnly = 1
WITH DOACTE AS(
SELECT ROW_NUMBER() OVER(PARTITION BY [File].Id ORDER BY CustomRecordsetId DESC) AS RowNumber, [File].*, FileType2Lexicon.Label as FileTypeLabel, [People].DefaultPhone, [People].InvertedName, CustomField.Name as FieldLabel, CustomFieldValue.Value as FieldValue
FROM FileType2Lexicon, SelectionCache, [People], [File]
INNER JOIN [CustomRecordSet]
ON [CustomRecordset].RecordId = [File].Id
INNER JOIN CustomFieldValue
ON [CustomRecordset].Id = CustomFieldValue.CustomRecordsetId
INNER JOIN [CustomField2Lexicon]
ON CustomField2Lexicon.CustomFieldId = CustomFieldValue.CustomFieldId
INNER JOIN [CustomField]
ON CustomField.Id = CustomField2Lexicon.CustomFieldId
WHERE [File].Id = SelectionCache.RecordId
AND SelectionCache.UUID = @UUID
AND SelectionCache.TableCode = @TableCode -- this is the code for File table
AND [File].Id <> 0
AND [File].FileTypeId = FileType2Lexicon.FileTypeId
AND FileType2Lexicon.LexiconId = @LexiconId
AND [File].ClientIdString = [People].ClientIdString
AND CustomFieldValue.Value <> ''),
SolicitorCTE AS(
SELECT [People].Name AS SolicitorName, [File].Id
FROM SelectionCache, [File]
INNER JOIN [People2File]
ON [People2File].FileId = [File].Id
INNER JOIN [Role2Lexicon]
ON [Role2Lexicon].RoleId = [People2File].RoleId
INNER JOIN [People]
ON [People].Id = [People2File].PeopleId
WHERE
[File].Id = SelectionCache.RecordId
AND SelectionCache.UUID = @UUID
AND SelectionCache.TableCode = @TableCode -- this is the code for File table
AND [File].Id <> 0
AND [Role2Lexicon].Label = 'Solicitor'),
ArrestingOfficerCTE AS(
SELECT ROW_NUMBER() OVER(PARTITION BY [File].Id ORDER BY [People].InvertedName ASC) AS RowNumber, [People].Name AS ArrestingOfficerName, [People].CompanyName AS ArrestingOfficerCompany, [File].Id
FROM SelectionCache, [File]
INNER JOIN [People2File]
ON [People2File].FileId = [File].Id
INNER JOIN [Role2Lexicon]
ON [Role2Lexicon].RoleId = [People2File].RoleId
INNER JOIN [People]
ON [People].Id = [People2File].PeopleId
WHERE
[File].Id = SelectionCache.RecordId
AND SelectionCache.UUID = @UUID
AND SelectionCache.TableCode = @TableCode -- this is the code for File table
AND [File].Id <> 0
AND [Role2Lexicon].Label = 'Arresting Officer'),
PivotCTE AS(
SELECT *
FROM
(Select Id, FieldLabel, FieldValue FROM DOACTE) AS Source
PIVOT(
MAX(FieldValue) FOR FieldLabel IN ([Date_Arrest], [Graphic_Client], [Ticket_1], [Ticket_2], [Ticket_3], [Ticket_4], [Ticket_5], [Charge_1], [Charge_2], [Charge_3], [Charge_4], [Charge_5])) as Pvt
)
SELECT DOACTE.*, COALESCE(ArrestingOfficerCTE.ArrestingOfficerCompany, 'NULL')AS ArrestingOfficerCompany, COALESCE(ArrestingOfficerCTE.ArrestingOfficerName, 'NULL') AS ArrestingOfficerName, SolicitorCTE.SolicitorName, PivotCTE.[Date_Arrest], dbo.GetImagebyId(PivotCTE.[Graphic_Client]) as Photo, PivotCTE.[Ticket_1], PivotCTE.[Ticket_2], PivotCTE.[Ticket_3], PivotCTE.[Ticket_4], PivotCTE.[Ticket_5], PivotCTE.[Charge_1], PivotCTE.[Charge_2], PivotCTE.[Charge_3], PivotCTE.[Charge_4], PivotCTE.[Charge_5]
FROM DOACTE
INNER JOIN
PivotCTE
ON DOACTE.Id = PivotCTE.Id
INNER JOIN
SolicitorCTE
ON DOACTE.Id = SolicitorCTE.Id
INNER JOIN
ArrestingOfficerCTE
ON DOACTE.Id = ArrestingOfficerCTE.Id
WHERE DOACTE.RowNumber = 1
AND ArrestingOfficerCTE.RowNumber = 1
ELSE
DELETE SelectionCache
WHERE UUID = @UUID
AND TableCode = @TableCode
END
答案 0 :(得分：1)
您可以保留联接以获取可选结果。我还需要为Arresting官员添加一个空检查到caluse的位置，这样即使是左连接，也不会排除记录。
SELECT DOACTE.*, COALESCE(ArrestingOfficerCTE.ArrestingOfficerCompany, 'NULL')AS ArrestingOfficerCompany, COALESCE(ArrestingOfficerCTE.ArrestingOfficerName, 'NULL') AS ArrestingOfficerName, SolicitorCTE.SolicitorName, PivotCTE.[Date_Arrest], dbo.GetImagebyId(PivotCTE.[Graphic_Client]) as Photo, PivotCTE.[Ticket_1], PivotCTE.[Ticket_2], PivotCTE.[Ticket_3], PivotCTE.[Ticket_4], PivotCTE.[Ticket_5], PivotCTE.[Charge_1], PivotCTE.[Charge_2], PivotCTE.[Charge_3], PivotCTE.[Charge_4], PivotCTE.[Charge_5]
FROM DOACTE
INNER JOIN
PivotCTE
ON DOACTE.Id = PivotCTE.Id
LEFT JOIN
SolicitorCTE
ON DOACTE.Id = SolicitorCTE.Id
LEFT JOIN
ArrestingOfficerCTE
ON DOACTE.Id = ArrestingOfficerCTE.Id
WHERE DOACTE.RowNumber = 1
AND ( ArrestingOfficerCTE.RowNumber = 1 or ArrestingOfficerCTE.RowNumber is null )
答案 1 :(得分：0)
你可以尝试：
SELECT DOACTE.*, COALESCE(ArrestingOfficerCTE.ArrestingOfficerCompany, 'NULL')AS ArrestingOfficerCompany, COALESCE(ArrestingOfficerCTE.ArrestingOfficerName, 'NULL') AS ArrestingOfficerName, SolicitorCTE.SolicitorName, PivotCTE.[Date_Arrest], dbo.GetImagebyId(PivotCTE.[Graphic_Client]) as Photo, PivotCTE.[Ticket_1], PivotCTE.[Ticket_2], PivotCTE.[Ticket_3], PivotCTE.[Ticket_4], PivotCTE.[Ticket_5], PivotCTE.[Charge_1], PivotCTE.[Charge_2], PivotCTE.[Charge_3], PivotCTE.[Charge_4], PivotCTE.[Charge_5]
into #Results
FROM DOACTE
INNER JOIN
PivotCTE
ON DOACTE.Id = PivotCTE.Id
INNER JOIN
SolicitorCTE
ON DOACTE.Id = SolicitorCTE.Id
INNER JOIN
ArrestingOfficerCTE
ON DOACTE.Id = ArrestingOfficerCTE.Id
WHERE DOACTE.RowNumber = 1
AND ArrestingOfficerCTE.RowNumber = 1
if ((select cout(*) from #Results) = 0)
begin
insert into #Results
select
.......your default values......
end
select * from #Results
drop table #Results
答案 2 :(得分：0)
这几乎是谎言19的回答，所以我在他第一次到达那里后就投了赞成票。我认为答案归结为LEFT JOIN你所在的INNER。由于我重新编写了您的查询，因此我实际上可以理解发生了什么，并且我做了一些您可能想要或可能不想要的更改，所以我会留在这里，以防它有用。
我的线路比你的更多，因为我的格式不同，部分是因为我使用临时表做了4次你做的事情。关于CTE的是，当你像这样堆叠它们时，你可能会得到一些非常粗糙的表现。 CTE的行为基本上就像一个即时视图，他们的SQL几乎在所有地方都重复出现。我正在限制[File]并将其移动到临时表中。然后我将它转换为另一个（因为这进一步限制了基于RowNumber为1的内容。然后我只使用最后的2个CTE并将其用于最终结果。
最后，SELECT * ...是邪恶的。我只是在寻找某些东西时才使用它。将它放入像这样的sproc会导致你的问题，如架构或CTE等等的变化。当你制作一个sproc，命名每一列时，它将在未来咬你更少。
如果您觉得它没有用，请随时忽略：
ALTER PROCEDURE [dbo].[GetRollCallData]
@Ids VARCHAR(255),
@LexiconId INT,
@UUID UNIQUEIDENTIFIER,
@ReadOnly INT
AS
BEGIN
DECLARE @TableCode INT = 58;
IF @Ids <> ''
BEGIN
EXEC InsertInSelectionCache @Ids, @UUID, @TableCode, 0
IF @ReadOnly = 1
BEGIN
SELECT
-- Add any other fully qualified columns. * is evil for sprocs. Only good for poking around in your own queries.
AllFiles.Id
INTO #SelectionFile
FROM
[File] AllFiles
INNER JOIN SelectionCache Cache
ON Cache.RecordId = AllFiles.Id
AND Cache.UUID = @UUID
AND Cache.TableCode = @TableCode -- this is the code for File table
WHERE
AllFiles.Id <> 0
;WITH DOACTE AS
(
SELECT
ROW_NUMBER() OVER(PARTITION BY SelectionFile.Id ORDER BY CustomRecordsetId DESC) AS RowNumber
, SelectionFile.Id AS FileId
-- [Replace with any other columns that you selected from [File] at the top.]
, Lexicon.Label AS FileTypeLabel
, [People].DefaultPhone
, [People].InvertedName
, CustomField.Name AS FieldLabel
, CustomFieldValue.Value AS FieldValue
FROM
#SelectionFile SelectionFile
INNER JOIN FileType2Lexicon Lexicon
ON Lexicon.FileTypeId = SelectionFile.FileTypeId
AND Lexicon.LexiconId = @LexiconId
INNER JOIN [People]
ON [People].ClientIdString = SelectionFile.ClientIdString
INNER JOIN [CustomRecordSet]
ON [CustomRecordset].RecordId = SelectionFile.Id
INNER JOIN CustomFieldValue
ON [CustomRecordset].Id = CustomFieldValue.CustomRecordsetId
AND CustomFieldValue.Value <> ''
INNER JOIN [CustomField2Lexicon]
ON CustomField2Lexicon.CustomFieldId = CustomFieldValue.CustomFieldId
INNER JOIN [CustomField]
ON CustomField.Id = CustomField2Lexicon.CustomFieldId
)
SELECT
Pvt.FileId
-- [Replace with any other columns that you selected from DOACTE you want to propagate to the end.]
, Pvt.[Date_Arrest]
, Pvt.[Graphic_Client]
, Pvt.[Ticket_1]
, Pvt.[Ticket_2]
, Pvt.[Ticket_3]
, Pvt.[Ticket_4]
, Pvt.[Ticket_5]
, Pvt.[Charge_1]
, Pvt.[Charge_2]
, Pvt.[Charge_3]
, Pvt.[Charge_4]
, Pvt.[Charge_5]
INTO #PivotedFile
FROM
(
SELECT
FileId
-- [Replace with any other columns that you selected from DOACTE you want to propagate to the end.]
, FieldLabel
, FieldValue
, FieldLabel AS PivotFieldLabel
, FieldValue AS PivotFieldValue
FROM
DOACTE
WHERE
RowNumber = 1
) AS Source
PIVOT
(
MAX(PivotFieldValue)
FOR PivotFieldLabel
IN
(
[Date_Arrest]
, [Graphic_Client]
, [Ticket_1]
, [Ticket_2]
, [Ticket_3]
, [Ticket_4]
, [Ticket_5]
, [Charge_1]
, [Charge_2]
, [Charge_3]
, [Charge_4]
, [Charge_5]
)
) as Pvt
;WITH SolicitorCTE AS
(
SELECT
[People].Name AS SolicitorName
, SelectedFiles.FileId
FROM
#PivotedFile SelectedFiles
INNER JOIN [People2File]
ON [People2File].FileId = SelectedFiles.FileId
INNER JOIN [Role2Lexicon]
ON [Role2Lexicon].RoleId = [People2File].RoleId
AND [Role2Lexicon].Label = 'Solicitor'
INNER JOIN [People]
ON [People].Id = [People2File].PeopleId
)
, ArrestingOfficerCTE AS
(
SELECT
ROW_NUMBER() OVER(PARTITION BY SelectionFile.Id ORDER BY [People].InvertedName ASC) AS RowNumber
, [People].Name AS ArrestingOfficerName
, [People].CompanyName AS ArrestingOfficerCompany
, SelectedFiles.FileId
FROM
#PivotedFile SelectedFiles
INNER JOIN [People2File]
ON [People2File].FileId = SelectedFiles.FileId
INNER JOIN [Role2Lexicon]
ON [Role2Lexicon].RoleId = [People2File].RoleId
AND [Role2Lexicon].Label = 'Arresting Officer'
INNER JOIN [People]
ON [People].Id = [People2File].PeopleId
)
SELECT
SelectedFiles.Id
-- [Replace with any other columns that you selected from DOACTE you want to propagate to the end.]
, COALESCE(ArrestingOfficerCTE.ArrestingOfficerCompany, 'NULL') AS ArrestingOfficerCompany
, COALESCE(ArrestingOfficerCTE.ArrestingOfficerName, 'NULL') AS ArrestingOfficerName
, SolicitorCTE.SolicitorName
, SelectedFiles.[Date_Arrest]
, dbo.GetImagebyId(SelectedFiles.[Graphic_Client]) as Photo
, SelectedFiles.[Ticket_1]
, SelectedFiles.[Ticket_2]
, SelectedFiles.[Ticket_3]
, SelectedFiles.[Ticket_4]
, SelectedFiles.[Ticket_5]
, SelectedFiles.[Charge_1]
, SelectedFiles.[Charge_2]
, SelectedFiles.[Charge_3]
, SelectedFiles.[Charge_4]
, SelectedFiles.[Charge_5]
FROM #PivotedFile SelectedFiles
LEFT JOIN SolicitorCTE
ON SelectedFiles.FileId = SolicitorCTE.FileId
LEFT JOIN ArrestingOfficerCTE
ON SelectedFiles.FileId = ArrestingOfficerCTE.FileId
AND ArrestingOfficerCTE.RowNumber = 1
DROP TABLE #SelectionFile
DROP TABLE #PivotedFile
END
ELSE
DELETE SelectionCache
WHERE
UUID = @UUID
AND TableCode = @TableCode
END
END
（如果您使用它，请阅读-- [Replace...]评论。