| Date: | Tue, 12 Oct 1999 22:31:15 GMT |
| Reply-To: | Bill Kossack <kossack@NETCOM1.NETCOM.COM> |
| Sender: | "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU> |
| From: | Bill Kossack <kossack@NETCOM1.NETCOM.COM> |
| Organization: | Netcom |
| Subject: | Re: Filtering out a phone number |
|---|
The best pattern I can see in the data is XXX XXX XXXX to identify phone
numbers.
How would I code the 'keep' for that?
GORRELP1 <GORRELP1@westat.com> wrote:
: 90% of this problem is being able to describe and distinguish the patterns for
: the 'keep' and the 'drop' data. The more specific the details you can furnish
: the easier the task will be. For example, how long can 'keep' number strings
: be? Does a three-digit number sequence, bordered by spaces, uniquely signal the
: start of an unwanted phone number? Etc.
: Paul Gorrell
: gorrelp1@westat.com
: ____________________Reply Separator____________________
: Subject: Re: Filtering out a phone number
: Author: Bill Kossack <kossack@NETCOM11.NETCOM.COM>
: Date: 10/12/1999 9:09 PM
: The rest of the string can contain numbers such as equipment model
: numbers that are of interest.
: Richard Read Allen <rrallen@ibm.net> wrote:
: : Is it possible for the useful consistent information to contain the digits
: : 0-9? If not, then use compress(x,'0123456789'). If it can contain digits, it's
: : a more difficult problem.
: --
: William S. Kossack
: Westminster, Colorado
: kossack@netcom.com
--
William S. Kossack
Westminster, Colorado
kossack@netcom.com
|