Question

我有一张发票表，其中包含InvoiceID（int identity）主键。还有一个InvoiceNumber列，它是一个整数。我有另一张表用于生成名为Invoice_Numbers的发票号码（见下文）。为了确保发票号码是唯一的并防止出现差距，我已经实施了以下代码。有人可以审查此代码并评论其可靠性。在多用户环境中运行SQL 2008。

在插入发票期间，用户在调用时获得相同的发票号码有多大的机会？

 
IF EXISTS
   (SELECT *
    FROM   sys.objects
    WHERE  object_id = OBJECT_ID(N'[Imports].[Invoices_Numbers]')
           AND type IN ( N'U' ))
  DROP TABLE [Imports].[Invoices_Numbers]

GO

CREATE TABLE [Imports].[Invoices_Numbers]
  (
     [InvoiceNumber] [INT] IDENTITY(1, 1) NOT NULL
     ,[Deleted]      [BIT] NOT NULL
     ,[Used]         [BIT] NOT NULL,
     CONSTRAINT [PK_Invoices_Numbers] PRIMARY KEY CLUSTERED ( [InvoiceNumber] ASC )WITH (PAD_INDEX = OFF, STATISTICS_NORECOMPUTE = OFF, IGNORE_DUP_KEY = OFF, ALLOW_ROW_LOCKS = ON, ALLOW_PAGE_LOCKS = ON) ON [PRIMARY]
  )
ON [PRIMARY]

GO

ALTER PROCEDURE [Imports].[Get_Invoice_Number]
(
  @InvoiceNumber INT OUTPUT
)
AS
  BEGIN
      DECLARE @NewNumber INT
      DECLARE @MinNumber INT

  BEGIN TRAN

  SELECT @MinNumber = MIN(InvoiceNumber)
  FROM   Imports.Invoices_Numbers

  IF @MinNumber > 1
    BEGIN
        SET IDENTITY_INSERT Imports.Invoices_Numbers ON;

        INSERT Imports.Invoices_Numbers
        (
          Invoicenumber
          ,Deleted
          ,Used
        )
        VALUES
        (
          1
          ,0
          ,1
        )
        SET IDENTITY_INSERT Imports.Invoices_Numbers OFF;
        SET @NewNumber=1
    END
  ELSE
    BEGIN
        WITH Gaps
             AS (SELECT TOP 1 a.InvoiceNumber + 1 AS GapValue
                 FROM   Imports.Invoices_Numbers a
                 WHERE  NOT EXISTS
                            (SELECT *
                             FROM   Imports.Invoices_Numbers b
                             WHERE  b.InvoiceNumber = a.InvoiceNumber + 1)
                        AND a.InvoiceNumber <
                            (SELECT MAX(InvoiceNumber)
                             FROM   Imports.Invoices_Numbers))
        SELECT @NewNumber = GapValue
        FROM   Gaps

        IF @NewNumber IS NULL
          BEGIN
              SELECT TOP 1 @NewNumber = InvoiceNumber
              FROM   Imports.Invoices_Numbers
              WHERE  Used = 0
                     AND Deleted = 0
              ORDER  BY InvoiceNumber

              IF @NewNumber IS NULL
                BEGIN
                    INSERT Imports.Invoices_Numbers
                    (
                      Deleted
                      ,Used
                    )
                    VALUES
                    (
                      0
                      ,1
                    )
                    SELECT @NewNumber = SCOPE_IDENTITY ()
                END
              ELSE
                BEGIN
                    UPDATE Imports.Invoices_Numbers
                    SET    Used = 1
                    WHERE  InvoiceNumber = @NewNumber
                END
          END
        ELSE
          BEGIN
              SET IDENTITY_INSERT Imports.Invoices_Numbers ON;

              INSERT Imports.Invoices_Numbers
              (
                Invoicenumber
                ,Deleted
                ,Used
              )
              VALUES
              (
                @NewNumber
                ,0
                ,1
              )
              SET IDENTITY_INSERT Imports.Invoices_Numbers OFF;
          END
    END

  SELECT @InvoiceNumber = @NewNumber

  COMMIT TRAN

END

Answer 1

您的解决方案看起来有点复杂。除了数据导入任务之外，我不建议使用set identity_insert。

为了确保唯一性，我首先添加一个唯一约束：

alter table Invoices add constraint UX_Invoices_InvoiceNr unique

然后你可以使用这样的SQL语句来插入发票：

while 1=1
    begin        
    declare @new_nr int
    select  @new_nr = max(InvoiceNr) + 1
    from    dbo.Invoices

    if @new_nr is null
        set @new_nr = 1

    insert  dbo.Invoices
            (InvoiceNr, ...)
    values  (@new_nr, ...)

    if @@rowcount = 1
        break
    end

顺序发票号码SQL服务器

1 个答案: