8

I have this long string in one of the columns of the table. I want to get only specific information:- My Table structure:-

Col1 = '123'
Col2 = 'AAAAA'
Col3 = 'Clent ID = 4356hy|Client Name = B B BOB|Client Phone = 667-444-2626|Client Fax = 666-666-0151|Info = INF8888877 -MAC333330554/444400800'

My select statement is:-

Select col1, col2, col3 from Table01

But in Col3 I just need 'Client Name's value which is 'B B BOB'.

In Col3 -

  • Column delimiter is '|' pipe char (eg. 'Client ID = 4356hy')

  • Key Value delimiter is ' = ' equal to sign with one white space (leading and trailing).

Please help.

2
  • 17
    Be prepared for a lot of comments on how terrible that design is :) Commented Feb 23, 2011 at 20:28
  • 1
    To pick up @JNK's ball: THAT'S A TERRIBLE DESIGN! But you can still save it, if you change the table structure to include just a foreign key to a client table that holds all that data... Commented Feb 23, 2011 at 20:36

4 Answers 4

9

For your specific data, you can use

Select col1, col2, LTRIM(RTRIM(SUBSTRING(
    STUFF(col3, CHARINDEX('|', col3,
    PATINDEX('%|Client Name =%', col3) + 14), 1000, ''),
    PATINDEX('%|Client Name =%', col3) + 14, 1000))) col3
from Table01

EDIT - charindex vs patindex

Test

select col3='Clent ID = 4356hy|Client Name = B B BOB|Client Phone = 667-444-2626|Client Fax = 666-666-0151|Info = INF8888877 -MAC333330554/444400800'
into t1m
from master..spt_values a
cross join master..spt_values b
where a.number < 100
-- (711704 row(s) affected)

set statistics time on

dbcc dropcleanbuffers
dbcc freeproccache
select a=CHARINDEX('|Client Name =', col3) into #tmp1 from t1m
drop table #tmp1

dbcc dropcleanbuffers
dbcc freeproccache
select a=PATINDEX('%|Client Name =%', col3) into #tmp2 from t1m
drop table #tmp2

set statistics time off

Timings

CHARINDEX:

 SQL Server Execution Times (1):
   CPU time = 5656 ms,  elapsed time = 6418 ms.
 SQL Server Execution Times (2):
   CPU time = 5813 ms,  elapsed time = 6114 ms.
 SQL Server Execution Times (3):
   CPU time = 5672 ms,  elapsed time = 6108 ms.

PATINDEX:

 SQL Server Execution Times (1):
   CPU time = 5906 ms,  elapsed time = 6296 ms.
 SQL Server Execution Times (2):
   CPU time = 5860 ms,  elapsed time = 6404 ms.
 SQL Server Execution Times (3):
   CPU time = 6109 ms,  elapsed time = 6301 ms.

Conclusion

The timings for CharIndex and PatIndex for 700k calls are within 3.5% of each other, so I don't think it would matter whichever is used. I use them interchangeably when both can work.

Sign up to request clarification or add additional context in comments.

1 Comment

Isn't PATINDEX('%|Client Name =%', col3) the same as CHARINDEX('|Client Name =', col3) and isn't the latter supposed to be faster than the former?
3

You need a split function:

SET ANSI_NULLS ON
GO
SET QUOTED_IDENTIFIER ON
GO
Create Function [dbo].[udf_Split]
(   
    @DelimitedList nvarchar(max)
    , @Delimiter nvarchar(2) = ','
)
RETURNS TABLE 
AS
RETURN 
    (
    With CorrectedList As
        (
        Select Case When Left(@DelimitedList, Len(@Delimiter)) <> @Delimiter Then @Delimiter Else '' End
            + @DelimitedList
            + Case When Right(@DelimitedList, Len(@Delimiter)) <> @Delimiter Then @Delimiter Else '' End
            As List
            , Len(@Delimiter) As DelimiterLen
        )
        , Numbers As 
        (
        Select TOP( Coalesce(DataLength(@DelimitedList)/2,0) ) Row_Number() Over ( Order By c1.object_id ) As Value
        From sys.columns As c1
            Cross Join sys.columns As c2
        )
    Select CharIndex(@Delimiter, CL.list, N.Value) + CL.DelimiterLen As Position
        , Substring (
                    CL.List
                    , CharIndex(@Delimiter, CL.list, N.Value) + CL.DelimiterLen     
                    , CharIndex(@Delimiter, CL.list, N.Value + 1)                           
                        - ( CharIndex(@Delimiter, CL.list, N.Value) + CL.DelimiterLen ) 
                    ) As Value
    From CorrectedList As CL
        Cross Join Numbers As N
    Where N.Value <= DataLength(CL.List) / 2
        And Substring(CL.List, N.Value, CL.DelimiterLen) = @Delimiter
    )

With your split function, you would then use Cross Apply to get the data:

Select T.Col1, T.Col2
    , Substring( Z.Value, 1, Charindex(' = ', Z.Value) - 1 ) As AttributeName
    , Substring( Z.Value, Charindex(' = ', Z.Value) + 1, Len(Z.Value) ) As Value
From Table01 As T
    Cross Apply dbo.udf_Split( T.Col3, '|' ) As Z

1 Comment

This function doesn't give consistent results. See stackoverflow.com/q/12480321 for issues, including a fiddle that demonstrates the problem.
0

You simply need to do a SUBSTR on the string in col3....

    Select col1, col2, REPLACE(substr(col3, instr(col3, 'Client Name'), 
    (instr(col3, '|', instr(col3, 'Client Name')  -
    instr(col3, 'Client Name'))
    ),
'Client Name = ',
'')
    from Table01 

And yes, that is a bad DB design for the reasons stated in the original issue

Comments

0

It is terrible, but you can try to use

select
SUBSTRING(Table1.Col1,0,PATINDEX('%|%=',Table1.Col1)) as myString
from
Table1

This code is probably not 100% right though. need to be adjusted

Comments

Start asking to get answers

Find the answer to your question by asking.

Ask question

Explore related questions

See similar questions with these tags.