from Master Ian, Praise indeed ! many thanks.

It may point up to those who have had to write then re-read data, that the update-able _infile_ might offer much potential for shorter solutions than used to be necessary.

Input statements provide so many text parsing capabilities !

With update-able _infile_, we can take advantage of these, without having to go to external text files.

(sounds like it might be one for coder's corner ! ) (.............but maybe a bit late with v9 around the corner )

Kind Regards Peter Crawford

for those who received this window in a useless wrapped state, here (in 70 column width)is the most important area, the 9 parts extracted from the input strings +FSVIEW: WORK.NEW_SCAN (B)------------------------------------------+ | Obs part1 part2 part3 part4 part5 part6 part7 part8 part9 | | | | 1 AAA BBB CC DD EE FF GG HHH | | 2 AAA HHH | | 3 AAA GGG | | 4 HHH | | 5 dlm BBB CCC / , FF GGG HHH | | 6 AAA H | | | +--------------------------------------------------------------------+

Datum: 01/05/2002 15:14 An: Peter Crawford/Zentrale/DeuBaExt@Zentrale

Betreff: RE: Function SCAN and delimiter question Nachrichtentext:

Peter,

Really neat trick!

IanWhitlock@westat.com

-----Original Message----- From: Peter Crawford [mailto:peter.crawford@DB.COM] Sent: Wednesday, May 01, 2002 10:56 AM To: SAS-L@LISTSERV.UGA.EDU Subject: Re: Function SCAN and delimiter question

My suggestion about using an infile buffer was intended to tolerate those features DSD supports, like more than one type of delimiter and delimiters embedded in data, masked in quotes

Having given it a try, it seems to be controlable Here is a log for a revision of Mike Zdeb's data (more delimiters and some quotes) with my _infile_ scan() process 357 data old; 358 infile datalines truncover ; 359 input string $char40.; 360 list;datalines;

RULE: ----+----1----+----2----+----3----+----4----+----5----+-- 361 AAA/BBB/CC/DD,EE/FF/GG/HHH 362 AAA///,,//HHH 363 AAA//////GGG/ 364 ////////HHH 365 dlm/BBB CCC/"/"/","/FF/GGG/HHH 366 AAA/////,/H NOTE: The data set WORK.OLD has 6 observations and 1 variables. NOTE: DATA statement used: real time 0.13 seconds cpu time 0.04 seconds

367 ; 368 data new_scan; 369 infile './*' dsd dlm=dlm truncover; /* infile used for scan */ 370 dlm = ",/" || "09"x ; 371 set old; 372 input @1 @@; /* establish a _infile_ buffer */ 373 _infile_=string; /* replace _infile_ with string to scan with DSD */ 374 input (part1 - part9)($) @@ ; 375 run;

NOTE: The infile './*' is: /* the first file in the current directory */ File Name=e:\users\crawford\sas8\addoptn.sas, File List=e:\users\crawford\sas8\*,RECFM=V, LRECL=256

NOTE: 1 record was read from the infile './*'. The minimum record length was 37. The maximum record length was 37. NOTE: There were 6 observations read from the data set WORK.OLD. NOTE: The data set WORK.NEW_SCAN has 6 observations and 10 variables. NOTE: DATA statement used: real time 0.73 seconds cpu time 0.03 seconds

376 dm log 'fsv;formula ' fsv; /* produces this window (because I have customised the view formula) */ +FSVIEW: WORK.NEW_SCAN (B)------------------------------------------------------------------------- ---------+ | Obs string part1 part2 part3 part4 part5 part6 part7 part8 part9 | | | | 1 AAA/BBB/CC/DD,EE/FF/GG/HHH AAA BBB CC DD EE FF GG HHH | | 2 AAA///,,//HHH AAA HHH | | 3 AAA//////GGG/ AAA GGG | | 4 ////////HHH HHH | | 5 dlm/BBB CCC/"/"/","/FF/GGG/HHH dlm BBB CCC / , FF GGG HHH | | 6 AAA/////,/H AAA H | | | +--------------------------------------------------------------------------- ---------------------------------+

Of course, it would be more interesting to implement on _real_ data !

Regards Peter Crawford

Datum: 01/05/2002 13:51 An: Peter Crawford/Zentrale/DeuBaExt@Zentrale

Betreff: Re: Function SCAN and delimiter question Nachrichtentext:

Hi. I think this works OK...TRANWRD then SCAN

data old; infile datalines truncover; input string $50.; datalines; AAA/BBB///EEE/FFF//HHH AAA/BBB/CCC/DDD/EEE/FFF/GGG/HHH AAA///////HHH AAA//////GGG/ ////////HHH AAA/////// ; run;

data new; length temp $100; array part(10) $20; set old; temp = tranwrd(string,'/','/~'); j=0; do while (1=1); j+1; part(j) = scan(temp,j,'/'); if part(j) eq '' then leave; part(j) = compress(part(j),'~'); end; drop j temp; run;

proc print data=new; var string part1-part8; run;

Mike Zdeb New York State Department of Health ESP Tower - Room 1811 Albany, NY 12237 P/518-473-2855 F/630-604-1475

Peter Crawford <peter.crawford@D To: SAS-L@LISTSERV.UGA.EDU B.COM> cc: Sent by: "SAS(r) Subject: Re: Function SCAN and delimiter question Discussion" <SAS-L@LISTSERV.U GA.EDU>

05/01/02 06:06 AM Please respond to Peter Crawford

The behaviour wanted by RICH0850 <rich0850@AOL.COM> is probably available "real-soon-now" in v9 ! But here is an interim alternative ! it only requires sas v8 to update _infile_ I haven't tried this ...yet, but why not use that DSD option on our text string ? Prepare a spare infile statement which you only use for this purpose. It should have the dsd option and dlm=<varname> (where that variable <varname> has the delimiters you need to use) When you want to use scan() with dsd behaviour, issue an infile statement using your special infile , place the text you want to scan, into _infile_ use input statements as necessary to extract from your string (but keep a trailing @ or @@ ) If the rest of the data step is using infile/inputs then you need to take of current the _infile_ buffer and column position before swithing to the special infile.

Has anyone tried this kind of thing ?

Regards Peter Crawford

Datum: 01/05/2002 01:16 An: SAS-L@LISTSERV.UGA.EDU

Antwort an: RICH0850 <rich0850@AOL.COM>

Betreff: Function SCAN and delimiter question Nachrichtentext:

Dear Group:

I want to use function SCAN as defined in the documentation. Unfortunately, if you have multiple occurances of the delimiter you have specified occuring in sequence, SCAN treats them as a single occurance. I want SCAN to return me results similar to the DSD option of the INFILE statement where multiple occurances of the delimiter are interpreted as missing values between delimiters.

Any ideas or work arounds?

Thanks.

--Richard

--

Diese E-Mail enthält vertrauliche und/oder rechtlich geschützte Informationen. Wenn Sie nicht der richtige Adressat sind oder diese E-Mail irrtümlich erhalten haben, informieren Sie bitte sofort den Absender und vernichten Sie diese Mail. Das unerlaubte Kopieren sowie die unbefugte Weitergabe dieser Mail ist nicht gestattet.

This e-mail may contain confidential and/or privileged information. If you are not the intended recipient (or have received this e-mail in error) please notify the sender immediately and destroy this e-mail. Any unauthorized copying, disclosure or distribution of the material in this e-mail is strictly forbidden.