Pavlos,
Here's my attempt:
*-------------------- example -------------------------
version 11
clear
input id str244( cit_1 company_1)
1 "US6449348-B1 3COM CORP _THRE-Non-standard_" "3COM CORP"
2 "US2004257999-A1 CETACEA NETWORKS CORP _CETA-Non-standard_" "CETACEA
NETWORKS CORP"
3 "US5566180-A HEWLETT-PACKARD CO _HEWP_" "HEWLETT-PACKARD CO"
4 "US6215865-B1 E-TALK CORP _ETAL-Non-standard_" "E-TALK CORP"
6 "US5600312-A MOTOROLA INC _MOTI_" "MOTOROLA INC"
7 "CONRED ELECTRONICS LTD _CONR-Non-standard_ MURAKOSHI S" "CONRED
ELECTRONICS LTD"
8 "TEMIC TELEFUNKEN MICROELECTRONIC GMBH _TELE_ LEICHT G, SCHUCH B"
"TEMIC TELEFUNKEN MICROELECTRONIC GMBH"
9 "US3476883-A" ""
10 "US5136671-A AT& T BELL LAB _AMTT_" "AT& T BELL LAB"
11 "US5195132-A AMERICAN TELEPHONE& TELEGRAPH CO _AMTT_" "AMERICAN
TELEPHONE& TELEGRAPH CO"
12 "US5605491-A CHURCH& DWIGHT CO INC _CHUR-Non-standard_" "CHURCH&
DWIGHT CO INC"
13 "US6028656-A CAMBRIDGE RES& INSTR INC _CAMB-Non-standard_"
"CAMBRIDGE RES& INSTR INC"
14 "US6201832 DAEWOO ELECTRONICS CO LTD _DAEW-Non-standard_ CHOI B"
"DAEWOO ELECTRONICS CO LTD"
15 "US6238946 INT BUSINESS MACHINES CORP _IBMC_ ZIEGLER J F" "INT
BUSINESS MACHINES CORP"
16 "US6947529-B2 -- US761995" ""
end
compress
gen co2 = trim(regexs(2)) if regexm(cit_1,"^(US[0-9]+[^ ]*)*([^_]+)_")
assert company_1 == co2
list, noobs
*-------------------- example -------------------------
Robert
On Thu, Apr 8, 2010 at 12:56 PM, Nick Cox<n.j.cox@durham.ac.uk> wrote:
I'd take a look at -split-. The recipe doesn't look simple even then given that your company names may contain blanks.
Nick
n.j.cox@durham.ac.uk
Pavlos C. Symeou
I am experiencing some problems with a command I use to extract a part
of a string variable which I use to create another string variable. The
existing string variable is cit_1 and may contain (one or multiple
instances of any of) a patent number (e.g. "US6449348-B1"), a company
name (e.g. "3COM CORP"), a company abbreviation enclosed by "_" (e.g.
"_THRE-Non-standard_"), other text after the "_" (e.g. see id 8). My aim
is to extract the company name, which appears always before its
abbreviation and use it to create a new string variable company_1. I
used the following command, which however fails to account for different
forms of the cit_1 values and produces incorrect company names.
gen company_1 = regexs(2) if (regexm(cit_1, "([A-Z0-9]*[\-][A-Z0-9]*[
\-]*) *([A-Z0-9 ]*)( *)([\_])(.*)([\_])"))
I provide below the various forms that cit_1 takes and how company_1
should look.
id cit_1 company_1
1 US6449348-B1 3COM CORP _THRE-Non-standard_ 3COM CORP
2 US2004257999-A1 CETACEA NETWORKS CORP _CETA-Non-standard_ CETACEA
NETWORKS CORP
3 US5566180-A HEWLETT-PACKARD CO _HEWP_ HEWLETT-PACKARD CO
4 US6215865-B1 E-TALK CORP _ETAL-Non-standard_ E-TALK CORP
US4528422-A -- US452232-A1 INTELEPLEX CORP _INTE-Non-standard_
INTELEPLEX CORP
6 US5600312-A MOTOROLA INC _MOTI_ MOTOROLA INC
7 CONRED ELECTRONICS LTD _CONR-Non-standard_ MURAKOSHI S CONRED
ELECTRONICS LTD
8 TEMIC TELEFUNKEN MICROELECTRONIC GMBH _TELE_ LEICHT G, SCHUCH B
TEMIC TELEFUNKEN MICROELECTRONIC GMBH
9 US3476883-A
10 US5136671-A AT& T BELL LAB _AMTT_ AT& T BELL LAB
11 US5195132-A AMERICAN TELEPHONE& TELEGRAPH CO _AMTT_ AMERICAN
TELEPHONE& TELEGRAPH CO
12 US5605491-A CHURCH& DWIGHT CO INC _CHUR-Non-standard_ CHURCH&
DWIGHT CO INC
13 US6028656-A CAMBRIDGE RES& INSTR INC _CAMB-Non-standard_ CAMBRIDGE
RES& INSTR INC
14 US6201832 DAEWOO ELECTRONICS CO LTD _DAEW-Non-standard_ CHOI B
DAEWOO ELECTRONICS CO LTD
15 US6238946 INT BUSINESS MACHINES CORP _IBMC_ ZIEGLER J F INT
BUSINESS MACHINES CORP
16 US6947529-B2 -- US761995
*
* 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/
*
* 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/
.
