topic Re: Parse text string for "Wyoming" or "WY" from txt file in SAS Programming

Parse text string for "Wyoming" or "WY" from txt file

Thalo — Thu, 06 Jun 2013 21:27:04 GMT

Hello I would like to get all the references from a text file in Var2 that match Wyoming or WY. I did right a routine that grabs WY specific items in VAR1 that substr(var1, 5, 7) = "WY";

This will grab most but not all of the files are consistent. So I would like to match things from the text field with the words wyoming, wy and etc. See example. And no Wyoming does not have a consistent place or structure in the text files. Sometimes it is middle and sometimes at the end of Var2.

Var1	Var2	Var3	VAR4	Var5	Last Updated
A\U\AUD10MD156N.csv	10-Month London Interbank Offered Rate (LIBOR), based on Australian Dollar	%	D	NSA	2013-05-15
G\B\GBP10MD156N.csv	10-Month London Interbank Offered Rate (LIBOR), based on British Pound	%	D	NSA	2013-05-15
C\A\CAD10MD156N.csv	10-Month London Interbank Offered Rate (LIBOR), based on Canadian Dollar	%	D	NSA	2013-05-15
D\K\DKK10MD156N.csv	10-Month London Interbank Offered Rate (LIBOR), based on Danish Krone (DISCONTINUED SERIES)	%	D	NSA	2013-04-10
E\U\EUR10MD156N.csv	10-Month London Interbank Offered Rate (LIBOR), based on Euro	%	D	NSA	2013-05-15
J\P\JPY10MD156N.csv	10-Month London Interbank Offered Rate (LIBOR), based on Japanese Yen	%	D	NSA	2013-05-15
N\Z\NZD10MD156N.csv	10-Month London Interbank Offered Rate (LIBOR), based on New Zealand Dollar (DISCONTINUED SERIES)	%	D	NSA	2013-03-07
S\E\SEK10MD156N.csv	10-Month London Interbank Offered Rate (LIBOR), based on Swedish Krona (DISCONTINUED SERIES)	%	D	NSA	2013-04-10
C\H\CHF10MD156N.csv	10-Month London Interbank Offered Rate (LIBOR), based on Swiss Frank	%	D	NSA	2013-05-15
U\USD10MD156N.csv	10-Month London Interbank Offered Rate (LIBOR), based on U.S. Dollar	%	D	NSA	2013-05-15
D\T\DTP10J22.csv	10-Year 0-1/8% Treasury Inflation-Indexed Note, Due 1/15/2022	%	D	NSA	2013-05-15
D\T\DTP10J23.csv	10-Year 0-1/8% Treasury Inflation-Indexed Note, Due 1/15/2023	%	D	NSA	2013-05-15
D\T\DTP10L22.csv	10-Year 0-1/8% Treasury Inflation-Indexed Note, Due 7/15/2022	%	D	NSA	2013-05-15
D\T\DTP10L21.csv	10-Year 0-5/8% Treasury Inflation-Indexed Note, Due 7/15/2021	%	D	NSA	2013-05-15
D\T\DTP10L20.csv	10-Year 1-1/4% Treasury Inflation-Indexed Note, Due 7/15/2020	%	D	NSA	2013-05-15
T\P\TP10L20.csv	10-Year 1-1/4% Treasury Inflation-Indexed Note, Due 7/15/2020	%	M	NSA	2013-05-01

Re: Parse text string for "Wyoming" or "WY" from txt file

PGStats — Thu, 06 Jun 2013 22:03:58 GMT

It is simple to do with Perl regular expression matching :

data wy;

set test;

if prxmatch("/\b(wy|wyoming)\b/io", var2);

run;

With your example data, it matches nothing.

Re: Parse text string for "Wyoming" or "WY" from txt file

ballardw — Thu, 06 Jun 2013 22:04:57 GMT

It may be that the function INDEXW is your friend here.

You didn't say how you want to indicate the presence but

data want;

set have;

WYflag = (indexw(upcase(var2),'WYOMING') >0 or indexw(upcaase(var2),'WY')>0);

run;

should create a flag variable =1 when either of the strings occur as a word ( for example showy wouldn't be counted as a match for wy) and 0 when neither is present in VAR2.

You may want to add delimiters to the INDEXW function if you data is likely to have something besides a space separating the words of interest.

Re: Parse text string for "Wyoming" or "WY" from txt file

Thalo — Thu, 06 Jun 2013 22:09:41 GMT

That works, I've not messed with prxmatch thanks,

Re: Parse text string for "Wyoming" or "WY" from txt file

Haikuo — Thu, 06 Jun 2013 22:09:41 GMT

PG, "/io" part of your code, I guess 'i' is to ignore case, is it? and what the 'o' do?

Thanks,

Haikuo

Re: Parse text string for "Wyoming" or "WY" from txt file

PGStats — Thu, 06 Jun 2013 22:20:44 GMT

Right. i requires to ignore the case, o requests that the pattern be compiled only once during the datastep execution (otherwise it is recompiled at every call). - PG

Re: Parse text string for "Wyoming" or "WY" from txt file

Haikuo — Thu, 06 Jun 2013 22:24:09 GMT

I see. I have been always wondering and always forgot asking. Thank you, PG!