<>
Diana:
Building on Matt Spittal's response, the function _regexs_ gives you
additional flexibility. For example,
input str15 name
"A.I.F. GMBH"
"A.I.F. COMPANY"
"QANTAS A.I.F."
"AIR NEW ZEALAND"
"Aero Dienst"
"Aer Arann"
end
generate str1 group_name = ""
replace group_name = regexs(1) if regexm(name,"[A-Za-z]*(A.I.F.)")
replace group_name = regexs(1) if regexm(name,"[A-Za-z]*(AIR)")
replace group_name = regexs(1) if regexm(name,"[A-Za-z]*(Aero*)")
list
+----------------------------+
| name group_~e |
|----------------------------|
1. | A.I.F. GMBH A.I.F. |
2. | A.I.F. COMPANY A.I.F. |
3. | QANTAS A.I.F. A.I.F. |
4. | AIR NEW ZEALAND AIR |
5. | Aero Dienst Aero |
6. | Aer Arann Aer |
+----------------------------+
Where the * in [A-Za-z]* and in (Aero*) means zero of more occurrences of
the previous character. The _replace_ commands could be incorporated in a
_foreach_ or _forvalues_ loop, or written as nested _cond_ functions.
Capitalization and punctuation could also be controlled for. But the above
is the basic idea.
..Frank
-----Original Message-----
From: [email protected]
[mailto:[email protected]] On Behalf Of Diana Eastman
Sent: Thursday, September 25, 2008 2:36 PM
To: [email protected]
Subject: st: Regexr Stata
Hi all,
I have a variable called "name" which lists several different airlines.
I need to write some code that will identify a regular expression within
these names and assign them a value in the new variable "group_name"
For instance, for the two names:
A.I.F. GMBH
A.I.F. COMPANY
I would want the group_name to be only "A.I.F." (the part of the string
they both share). The identifying string does not occur in the same
position across the names.
Any help is greatly appreciated.
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
