Date:         Thu, 6 Oct 2005 01:17:28 -0700
Reply-To:     Cady Wan <xqw1@yahoo.com>
Sender:       "SPSSX(r) Discussion" <SPSSX-L@LISTSERV.UGA.EDU>
From:         Cady Wan <xqw1@yahoo.com>
Subject:      Re: Extracting information from post codes.
In-Reply-To:  <5.1.0.14.2.20051005175458.0323ec10@pop.mindspring.com>
Content-Type: text/plain; charset=iso-8859-1

This also can be done by:

*##########################################.
data list list /current (a8).
begin data
DG10 8SR
M694JD
BT3 1DR
end data.
*##########################################.

loop #i = length(current) to 1 by -1.
- if index('0123456789',substr(current,#i,1))>0
firstdigit = #i.
end loop.
string required(a4).
do if (firstdigit > 1).
 compute required = substr(current, 1, firstdigit-1).
end if.
if (sysmis(firstdigit)) required = current.
*#############################################.

Regards,
Cady



--- Richard Ristow <wrristow@mindspring.com> wrote:

> At 01:34 PM 10/5/2005, David Lindsay wrote:
>
> >I have a number of GB post codes. [...] I guess
> that I need to detect
> >the first [digit] and then extract the letters to
> the left."
>
> Art's posted a fine solution. Here's an alternative,
> which I'm posting
> mainly to illustrate a 'cute' use of INDEX: Finding
> the first
> occurrence of any character on a list. (Alas,
> there's no way to find
> the first occurrence of any character NOT on a
> list.)
>
> If there's any practical advantage, it's that this
> code will pick off a
> string of any length preceding the first digit.
> Note, also, that '#IDX'
> is a scratch variable, so will not be present in
> (possibly, clutter)
> the output file. (On the other hand, if it were a
> normal variable, it
> would be easier to diagnose bugs.)
>
> *  ............................
> LIST.
> Notes
>
|---------------------------|-----------------------|
> |Output Created             |05 Oct 05 17:52:27
> |
>
|---------------------------|-----------------------|
> CURRENT    REQUIRED
>
> DG10 8SR   DG
> M694JD     M
> BT3 1DR    BT
>
> Number of cases read:  3    Number of cases listed:
> 3
>
>
> STRING  COMPUTED (A4).
> COMPUTE #IDX = INDEX(CURRENT,'0123456789',1).
> COMPUTE COMPUTED = SUBSTR(CURRENT,1,#IDX-1).
>
> LIST.
> Notes
>
|---------------------------|-----------------------|
> |Output Created             |05 Oct 05 17:52:27
> |
>
|---------------------------|-----------------------|
> CURRENT    REQUIRED COMPUTED
>
> DG10 8SR   DG       DG
> M694JD     M        M
> BT3 1DR    BT       BT
>
> Number of cases read:  3    Number of cases listed:
> 3
>





__________________________________
Yahoo! Mail - PC Magazine Editors' Choice 2005
http://mail.yahoo.com