如何将字符串传递给批量插入而不是文件？ [英] How to pass String to Bulk insert instead of file?

问题描述

我以前使用大容量插入命令来转换Csv文件的int表。现在我将CSV文件保存为VarBinary值在sql server中。现在我可以通过使用CAST和CONVERT函数将其类型转换为Varchar来从Varbinary文件中获取数据。但是现在我遇到一个问题，我无法使用批量插入将包含CSV内容的Varchar字符串转换为表。任何人都可以帮我
我的示例代码如下：

I used to use bulk insert command to Convert a Csv file int table.Resently i saved a CSV file as a VarBinary value in sql server.Now I can get data from Varbinary file by typecasting it to Varchar using CAST and CONVERT functions.But now i got an issue i cant convert this Varchar String containing csv content to table using bulk insert.Can any one help me My example code is given below:

--@String contains varchar value of CSV file content.  
SET @sql = 'BULK INSERT TempCsv
FROM  ''' + @String  + '''
WITH
(
    FIRSTROW = 2,
    FIELDTERMINATOR = '','',  
    ROWTERMINATOR = ''\n'',   
    TABLOCK
)'

请帮帮我，有什么方法或替代方法可以将csv字符串中的数据插入表中。

Please help me.Is there any way or alternative to insert data from csv string to table.

推荐答案

编辑：允许使用多个字符分隔符

allow multiple char separators

这就是我的方法解决了。它涉及：

This is how I solved it. It involves:

• 用于拆分换行符的表值函数（ xftSplit ） （char（10））插入表格行


• 用于提取行的第n个字段的标量函数（ fSubstrNth ），给定分隔符


• 用于查找分隔符的第n个索引的标量函数（ fPatIndexMulti ） b
• （可选）替代向右函数以接受负值


• 最后，在解决方案中使用一些特定的代码，因为SQL不允许动态表函数定义（换句话说，您不能从具有动态列的函数中进行 SELECT ）

• A table-valued function (xftSplit) for spliting line breaks (char(10)) into table lines
• A scalar function (fSubstrNth) for extracting the n-th field of a line, given an separator
• A scalar function (fPatIndexMulti) for finding the n-th index of the separator
• (Optional) alternative Right function to accept negative values
• Finally, some specific code to use in your solution, since SQL doesn't allow dynamic table-function definitions (in other words, you can't SELECT from a function with dynamic columns)

现在，用于代码段：

xftSplit

-- =============================================
-- Author:      Bernardo A. Dal Corno
-- Create date: 15/07/2014
-- Description: Quebra valores a partir de caracteres e retorna lista em tabela
-- =============================================
CREATE FUNCTION [dbo].[xftSplit] 
(
    @Texto varchar(max),
    @Splitter varchar(3)
)
RETURNS 
@Lista TABLE 
(
    ValoresQuebrados varchar(8000)
)
AS
BEGIN
DECLARE @Pos Smallint

  While len(@Texto)>0 BEGIN
    SET @Pos = Patindex('%'+@Splitter+'%',@Texto)

    IF @Pos > 0 BEGIN
      INSERT INTO @Lista
      SELECT left(@Texto, @Pos-1)

      SET @Texto = right(@Texto, len(@Texto)-@Pos)
    END
    ELSE BEGIN
      INSERT INTO @Lista
      SELECT @Texto

      SET @Texto = ''
    END
  End

  RETURN 
END

fSubstrNth

-- =============================================
-- Author:      Bernardo A. Dal Corno
-- Create date: 18/07/2017
-- Description: substring com 2 PatIndex limitando inicio e fim
-- =============================================
CREATE FUNCTION fSubstrNth
(
  @Text varchar(max),
  @Sep varchar(3),
  @N int --Nth campo
)
RETURNS varchar(max)
AS
BEGIN
  DECLARE @Result varchar(max)

  IF @N<1 RETURN ''
  IF @N=1
    SET @Result = substring(@Text, 1, dbo.fPatIndexMulti(@Sep,@Text,1)-1)
  ELSE
    SET @Result = substring(@Text, dbo.fPatIndexMulti(@Sep,@Text,@N-1)+LEN(@Sep), CASE WHEN dbo.fPatIndexMulti(@Sep,@Text,@N)>0 THEN dbo.fPatIndexMulti(@Sep,@Text,@N)-dbo.fPatIndexMulti(@Sep,@Text,@N-1)-LEN(@Sep) ELSE LEN(@Text)+1 END)

  RETURN @Result
END

fPatIndexMulti

-- =============================================
-- Author:      Bernardo A. Dal Corno
-- Create date: 17/07/2017
-- Description: recursive patIndex
-- =============================================
CREATE FUNCTION [dbo].[fPatIndexMulti]
(
  @Find varchar(max),
  @In varchar(max),
  @N tinyint
)
RETURNS int
AS
BEGIN
  DECLARE @lenFind int, @Result int, @Texto varchar(max), @index int
  DECLARE @i tinyint=1

  SET @lenFind = LEN(@Find)-1
  SET @Result = 0
  SET @Texto = @In
  WHILE (@i <= @N) BEGIN
    SET @index = patindex('%'+@Find+'%',@Texto)
      IF @index = 0 RETURN 0
    SET @Result = @Result + @index
    SET @Texto = dbo.xRight(@Texto, (@index + @lenFind)*-1)

    SET @i = @i + 1
  END
  SET @Result = @Result + @lenFind*(@i-2)

  RETURN @Result
END

xRight

-- =============================================
-- Author:      Bernardo A. Dal Corno
-- Create date: 06/01/2015
-- Description: Right inverso (para nros < 0)
-- =============================================
CREATE FUNCTION [dbo].[xRight] 
(
  @Texto varchar(8000),
  @Qntd int
)
RETURNS varchar(8000)
AS
BEGIN
  DECLARE @Result varchar(8000)

  IF (Len(@Texto) = 0) OR (@Qntd = 0)
    SET @Result = ''
  ELSE IF (@Qntd > 0) 
      SET @Result = Right(@Texto, @Qntd)
    ELSE IF (@Qntd < 0)
    SET @Result = Right(@Texto, Len(@Texto) + @Qntd)

  RETURN @Result
END

特定代码

SELECT 
     acolumn = 'any value',
     field1 = dbo.fSubstrNth(line,',',1),
     field2 = dbo.fSubstrNth(line,',',2),
     anothercolumn = 'set your query as you would normally do',
     field3 = (CASE dbo.fSubstrNth(line,',',3) WHEN 'C' THEN 1 ELSE 0 END)
FROM (
  SELECT line = ValoresQuebrados FROM dbo.xftSplit(@StringVariable, char(10))
) lines

请注意：

• fSubstrNth 接收要从行中提取的第n个字段


• xftSplit 接收到一个变量，该变量包含要从中进行批量存储的字符串（无论来源如何）以及 char（10）作为的分隔符\n ，但也可以是其他任何内容


• 查询可以与其他任何查询相同。这意味着它可以存储在过程，表函数，视图等中。您可以按所需的任何顺序提取一些或所有字段，并根据需要进行处理


• 如果使用在存储过程中，您可以创建一种创建查询和临时表的通用方法，该表将使用动态列加载字符串，但是您必须调用另一个过程以使用数据，或者像上面一样在上面创建特定的查询程序（这将使其变得非通用，只是可重用）

• fSubstrNth receives the n-th field to extract from the line
• xftSplit receives a variable that contains the string you want to bulk from (whatever the source) and a char(10) as the splitter of \n, but it could be anything else
• The query can be as any other. This means it can be stored in a procedure, tabled-function, view, etc. You can extract some or all fields, in any order you wish, and process however you want
• If used in a stored procedure, you could create a generic way of creating a query and temp table that loads the string with dynamic columns, but you have to make a call to another procedure to use the data OR create a specific query like above in the same procedure (which would make it non-generic, just more reusable)

这篇关于如何将字符串传递给批量插入而不是文件？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！