我需要编写一个输入为字符串的存储过程。
输入字符串包含变量名称,它们的值由管道分隔符分隔,如下所示:
Name =Praveen | City=Hyderabad | Mobile=48629387429| Role=User| etc
在存储过程中,我声明了@x, @y, @z, @t
之类的变量来获取值
@x=Praveen (Name value)
@y=Hyderabad (City Value)
@z=48629387429(Mobile Value)
@t=User(Role Value)
此外,输入字符串可以按任何顺序显示值,如
City=Hyderabad | Mobile=48629387429 | Role=User | Name =Praveen |etc
一旦我将值解析为@x, @y, @z, @t
等,我必须在存储过程中使用这些值。
请告诉我如何解析输入字符串以分别获取Name, City, Mobile, Role
到@x, @y, @z
和@t
的值。
答案 0 :(得分:4)
一种可能的解决方案是使用XML
DECLARE @text VARCHAR(1000)
,@xml xml
SELECT @text = 'City=Hyderabad | Mobile=48629387429 | Role=User | Name =Praveen'
SELECT @text = REPLACE(@text,'|','"')
,@text = REPLACE(@text,'=','="')
,@text = '<row ' + @text + '"/>'
SELECT @xml = CAST(@text AS XML)
select
line.col.value('@Name[1]', 'varchar(100)') AS Name
,line.col.value('@City[1]', 'varchar(100)') AS City
,line.col.value('@Mobile[1]', 'varchar(100)') AS Mobile
,line.col.value('@Role[1]', 'varchar(100)') AS Role
FROM @xml.nodes('/row') AS line(col)
答案 1 :(得分:0)
假设你的输入参数叫做@Text。
DECLARE @Text varchar(255),
@x varchar(255)
SET @Text = 'Name=Praveen | City=Hyderabad | Mobile=48629387429| Role=User'
-- Added to show how to account for non-trailing |
SET @Text = @Text + ' | ';
SET @x = LTRIM(RTRIM(substring(
@Text,
charindex('Name=', @Text) + LEN('Name='),
charindex(' | ', @Text, charindex('Name=', @Text)) - LEN('Name=')
)))
SELECT @x
然后只需重复一次@y,@ z,@ t将Name =改为你的休息时间。
答案 2 :(得分:0)
我绝对建议在程序端进行字符串解析而不是数据端。话虽这么说,如果你绝对必须尝试做类似的事情:
DECLARE @String [nvarchar](256) = 'Name=Praveen | City=Hyderabad | Mobile=48629387429 | Role=User |'
DECLARE @name [nvarchar](256) = (SELECT SUBSTRING(@String, CHARINDEX('Name=', @String)+5, CHARINDEX('|', @String)))
DECLARE @city [nvarchar](256) = (SELECT SUBSTRING(@String, CHARINDEX('City=', @String)+5, CHARINDEX('|', @String)))
DECLARE @mobile [nvarchar](256) = (SELECT SUBSTRING(@String, CHARINDEX('Mobile=', @String)+7, CHARINDEX('|', @String)))
DECLARE @role [nvarchar](256) = (SELECT SUBSTRING(@String, CHARINDEX('Role=', @String)+5, CHARINDEX('|', @String)))
SELECT RTRIM(LTRIM(LEFT(@name, CHARINDEX('|', @name)-1))) AS Name,
RTRIM(LTRIM(LEFT(@city, CHARINDEX('|', @city)-1))) AS City,
RTRIM(LTRIM(LEFT(@mobile, CHARINDEX('|', @mobile)-1))) AS Mobile,
RTRIM(LTRIM(LEFT(@role, CHARINDEX('|', @role)-1))) AS Role
返回:
Name | City | Mobile | Role
________________________________________________
Praveen | Hyderabad | 48629387429 | User
请注意,初始查询中CHARINDEX
添加的长度等于搜索字符串。
“Name =”等于5个字符,所以我们添加5来将索引移过=符号, “Mobile =”等于7,所以我们加7。
同样在最后SELECT
查询中,我们从每个CHARINDEX
中减去1以删除|符号
来源:
答案 3 :(得分:0)
这是一种使用循环进行字符串操作的有趣方式。请注意我们如何定义@x,@ y等变量以获取特定值。
-- Simulate proc parameter
declare @input nvarchar(max) = 'Name =Praveen | City=Hyderabad | Mobile=48629387429| Role=User'
-- OP's preferred destination vars
declare @x nvarchar(max) = 'Name'
declare @y nvarchar(max) = 'City'
declare @z nvarchar(max) = 'Mobile'
declare @t nvarchar(max) = 'Role'
-- The key/value delimiters we are expecting
declare @recordDelim nchar(1) = '|'
declare @valueDelim nchar(1) = '='
-- Temp storage
declare @inputTable table (
name nvarchar(128) not null primary key
, value nvarchar(max) null
)
-- Get all key/value pairs
while ltrim(rtrim(@input)) != '' begin
insert into @inputTable (name) select ltrim(rtrim(replace(left(@input, isnull(nullif(charindex(@recordDelim, @input), 0), len(@input))), @recordDelim, '')))
select @input = ltrim(rtrim(right(@input, len(@input) - isnull(nullif(charindex(@recordDelim, @input), 0), len(@input)))))
end
-- Separate keys and values
update @inputTable
set name = ltrim(rtrim(left(name, isnull(nullif(charindex(@valueDelim, name) - 1, 0), len(name)))))
, value = ltrim(rtrim(right(name, len(name) - isnull(nullif(charindex(@valueDelim, name), 0), len(name)))))
-- Populate the variables
-- If any are null, then this key/value wasn't present
set @x = (select value from @inputTable where name = @x)
set @y = (select value from @inputTable where name = @y)
set @z = (select value from @inputTable where name = @z)
set @t = (select value from @inputTable where name = @t)
另外,从您输入的不规则间距来看,我猜你想要修剪所有内容(这就是为什么这个proc会在整个地方做到这一点)。