I've been following this thread and thinking the same. Seems to make the code robust you need to consider issues with the data entry, unless you have a QA/QC program on the front end.

Warren Schlechte

-----Original Message----- From: toby dunn [mailto:tobydunn@HOTMAIL.COM] Sent: Mon 12/27/2010 8:05 PM Subject: Re: Is there an easier way to solve this?

Art,

Since you essentually dealing with a free form text string I have a question. Is there the possibility that any of the three parts you want to parse will be missing?

Toby Dunn

"I'm a hell bent 100% Texan til I die"

"Don't touch my Willie, I don't know you that well"

> Date: Mon, 27 Dec 2010 20:14:40 -0500 > From: art297@ROGERS.COM > Subject: Re: Is there an easier way to solve this? > To: SAS-L@LISTSERV.UGA.EDU > > Max, > > Tonight you get a free lesson. > > When you want to know if something is more or less efficient, DON'T ask the > list! Just build some code that allows you to make the two (or however > many) sets of code comparable and run them. > > If you find something that the rest of us might be interested, then share > the results. > > Thus, for your current question, wouldn't something like the following tell > you what you want to know?: > > data _null_; > file "c:\testdata.txt"; > input; > do i=1 to 100000; > put _infile_; > end; > cards; > Uptown, Smalltown XX 12345 > The City of New York Big Apple Big Apple NY 85044 > ; > run; > > data address; > length city $40 state $2 zip $5; > informat zip $revers5.; > informat state $revers2.; > informat city $revers40.; > infile "c:\testdata.txt"; > input @; > _infile_=reverse(_infile_); > input zip state city &; > run; > > data b; > Length zip $ 5 state $ 2 city $ 40; > infile "c:\testdata.txt"; > input; > Zip=scan(_infile_, -1); > State=scan(_infile_, -2); > City=tranwrd(_infile_, state||" "||zip, " "); > run; > > The log will tell you everything that you want to know and you, then, can > decide if it is just something to learn or something you should share. > > In the present case, it might just be worth sharing. > > Art > ------- > On Mon, 27 Dec 2010 18:26:35 -0500, bbser 2009 <bbser2009@GMAIL.COM> wrote: > > >Nat > > > > > >Yes. And alternative to my first code using tranwrd(), maybe it would be > >robust to adjust it like this (using tranwrd() twice instead of just once): > > > >... > >zip=scan(x, -1); > >temp=tranwrd(x, zip, " "); *First get rid of the value of zip from the full > >string in x; > >state=scan(temp, -1); *It is minus one, not minus two; > >city=tranwrd(temp, state, " "); *Secondly get rid of state from temp; > >... > > > >I guess this will get rid of the problem of varying number of blanks and is > >better than my second code where I used scan() and catx(). > >Now I am wondering, how this compares to Saren's in term of efficiency? > > > >Max > > > >-----Original Message----- > >From: Nat Wooding [mailto:nathani@verizon.net] > >Sent: December-27-10 5:28 PM > >To: 'bbser 2009' > >Subject: RE: [SAS-L] Is there an easier way to solve this? > > > >Max > > > >I don't use Tranwrd enough to know all of its nuances. If you used > > > >Record = compbl( record ); > > > >You would get rid of extra blanks. > > > >Nat > > > >-----Original Message----- > >From: bbser 2009 [mailto:bbser2009@gmail.com] > >Sent: Monday, December 27, 2010 5:04 PM > >To: 'Nat Wooding' > >Cc: SAS-L@LISTSERV.UGA.EDU > >Subject: RE: [SAS-L] Is there an easier way to solve this? > > > >Nat > > > >Thanks for let me know the "continue" statement. Glad to add it to my > >arsenal. > >As for my earlier code, i just thought it might not be robust or something. > >For example, if some of the records like below have two more spaces between > >NY and 111111. > > > >xxxx xxx xx NY 111111 > > > >Then using tranwrd(record, a||""||b, "") does not replace "NY 111111" > >totally with blanks. > > > >Max > > > >-----Original Message----- > >From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU] On Behalf Of Nat > >Wooding > >Sent: December-27-10 4:08 PM > >To: SAS-L@LISTSERV.UGA.EDU > >Subject: Re: [SAS-L] Is there an easier way to solve this? > > > >Max > > > >SAS fussed at me until I made the variable for City longer (I used 50). > > > >Your earlier solution was simpler but this does work. One thing that I > would > >do would be to stop the loop as soon as it got to the end of the words in > >the string as in the following code. > > > >Nat > > > >do i=1 to 20; > > word[i]=scan(x,-i); > > if word[i]='' then continue;**<<<< new line; > >end; > > > >-----Original Message----- > >From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU] On Behalf Of bbser > >2009 > >Sent: Monday, December 27, 2010 3:42 PM > >To: SAS-L@LISTSERV.UGA.EDU > >Subject: Re: Is there an easier way to solve this? > > > >Got typos in the code. Here is the new one. > >------------------ > >How about this below? It looks "elementary" for newbies like me and seemly > >works fine for whatever the longest USA city names. > > > >Max > > > >---------- > >data a; > > keep x zip state city; > > x="The City of New York Big Apple Big Apple NY 85044"; > > Length zip $ 5 state $ 2 city $ 30; > > array word{20} $ 15; > > do i=1 to 20; > > word[i]=scan(x,-i); > > end; > > zip=word[1]; > > state=word[2]; > > city=catx("", of word20-word3); > >run; > >proc print; > >run; > > > >-----Original Message----- > >From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU] On Behalf Of Nat > >Wooding > >Sent: December-27-10 1:16 PM > >To: SAS-L@LISTSERV.UGA.EDU > >Subject: Re: [SAS-L] Is there an easier way to solve this? > > > >Matt > > > >I took the liberty to send my reply to the list in case there are any bored > >Birdies listening in. > > > >I totally agree that SAS is not publicizing it as well as they could but I > >am generally unhappy with having the NLS formats and informats segregated > >from their peers, particularly since some are very useful. > > > >Did you notice the companion informat > > > >$REVERJw.@ inputs text right to left, preserves leading and trailing > >blanks > > ABCD | $reverj6. | ' DCBA' > > > >I copied this from TS486 and not the standard docs. > > > >Nat > >-----Original Message----- > >From: matt.pettis@thomsonreuters.com > [mailto:matt.pettis@thomsonreuters.com] > > > >Sent: Monday, December 27, 2010 12:43 PM > >To: nathani@VERIZON.NET > >Subject: RE: Is there an easier way to solve this? > > > >Thanks Nat! It is indeed a nice informat to keep in your back pocket... > >just think SAS shouldn't be hiding this light under a bushel... > > > >Thanks again, > >Matt > > > >-----Original Message----- > >From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU] On Behalf Of Nat > >Wooding > >Sent: Monday, December 27, 2010 11:38 AM > >To: SAS-L@LISTSERV.UGA.EDU > >Subject: Re: Is there an easier way to solve this? > > > >Matt > > > >Art and I spoke of this offline earlier today. 9.1.3 docs have an entry for > >it within the normal informats but refer you to the NLS docs. > > > >Art and I also corresponded about the width defaulting to 1. I find that if > >I have a length statement in the code, I do not need to supply a width. > > > >Nat > > > >-----Original Message----- > >From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU] On Behalf Of > Matthew > >Pettis > >Sent: Monday, December 27, 2010 12:28 PM > >To: SAS-L@LISTSERV.UGA.EDU > >Subject: Re: Is there an easier way to solve this? > > > >That '$revers.' informat wasn't documented in my local SAS Help files. I > >had to google it and found that it is a NLS informat. Anybody know why it > >wouldn't be in the base help files that come with my SAS install (and I > have > >9.2). > > > >Just curious, > >Thanks, > >matt > > > >-----Original Message----- > >From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU] On Behalf Of Arthur > >Tabachneck > >Sent: Monday, December 27, 2010 9:42 AM > >To: SAS-L@LISTSERV.UGA.EDU > >Subject: Re: Is there an easier way to solve this? > > > >Nat pointed out to me, offline, that I had left off the critical statement > >in Søren's code, namely the informat assignment: > > > > informat zip state city_soren $revers.; > > > >Yes, indeed, a VERY nice solution. > > > >Art > >-------- > >On Mon, 27 Dec 2010 10:05:50 -0500, Arthur Tabachneck <art297@ROGERS.COM> > >wrote: > > > >>Nat, > >> > >>Either you (or I) have had too much Excedrin or our systems are > functioning > >>differently (hmmm .. may they've had too much Excedrin). Running your > code > >>I get (what I expected), Søren's city 'k' for New York and Søren's city > >'a' > >>for Vienna. > >> > >>If you get something different, please email me a copy of the resulting > >>file. > >> > >>Art > >>-------- > >>On Mon, 27 Dec 2010 09:27:12 -0500, Nat Wooding <nathani@VERIZON.NET> > >wrote: > >> > >>>Art > >>> > >>>It looks to me that both solutions produce identical results. Try the > >>>following (which includes a merge sans by statement!!) > >>> > >>>Nat > >>> > >>> data soren; > >>> length firstname lastname address_SOREN $20 city_soren $40 state $2 zip > >>>$5; > >>> informat zip state city_soren $revers.; > >>> input firstname lastname / > >>> address_SOREN & / > >>> @; > >>> _infile_=reverse(_infile_); > >>> input zip state city_soren &; > >>> drop zip state firstname lastname; > >>>cards; > >>>Lee Athnos > >>>1215 Raintree Circle > >>>New York NY 85044 > >>>Heidie Baker > >>>1751 Diehl Road > >>>Vienna VA 22124 > >>>;run; > >>> > >>>data Art; > >>> Length City_art $ 40 > >>> State $ 2 > >>> Zip $ 5 > >>>; > >>>keep city_art address_art; > >>> infile cards ; > >>> input FirstName $ LastName $ / > >>> Address_art $ 1 - 20 / > >>> @; > >>> _infile_ = reverse(_infile_); > >>> input zip state city_art &; > >>> city_art=reverse(trim(city_art)); > >>> state=reverse(state); > >>> zip=reverse(zip); > >>> cards; > >>>Lee Athnos > >>>1215 Raintree Circle > >>>New York NY 85044 > >>>Heidie Baker > >>>1751 Diehl Road > >>>Vienna VA 22124 > >>>; > >>>run; > >>> > >>>Data Test; > >>>merge soren art; > >>>run; > >>> > >>> > >>>-----Original Message----- > >>>From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU] On Behalf Of > >Arthur > >>>Tabachneck > >>>Sent: Monday, December 27, 2010 9:16 AM > >>>To: SAS-L@LISTSERV.UGA.EDU > >>>Subject: Re: Is there an easier way to solve this? > >>> > >>>Søren, > >>> > >>>Definitely creates less of a dependency on Excedrin. However, it would > >>>require three more uses of the reverse function and, city appears strange > >>>unless one trims the leading(following?) space: > >>> > >>>data Address; > >>> Length City $ 40 > >>> State $ 2 > >>> Zip $ 5 > >>>; > >>> infile cards ; > >>> input FirstName $ LastName $ / > >>> Address $ 1 - 20 / > >>> @; > >>> _infile_ = reverse(_infile_); > >>> input zip state city &; > >>> city=reverse(trim(city)); > >>> state=reverse(state); > >>> zip=reverse(zip); > >>> cards; > >>>Lee Athnos > >>>1215 Raintree Circle > >>>New York NY 85044 > >>>Heidie Baker > >>>1751 Diehl Road > >>>Vienna VA 22124 > >>>; > >>>run; > >>> > >>>But, definitely a nice way to accomplish the task. > >>> > >>>Art > >>>-------- > >>>On Mon, 27 Dec 2010 01:11:50 -0500, S=?ISO-8859-1?Q?=C3=B8ren?= Lassen > >>><s.lassen@POST.TELE.DK> wrote: > >>> > >>>>Art, > >>>>How about this: > >>>>data address; > >>>> length firstname lastname address $20 city $40 state $2 zip $5; > >>>> informat zip state city $revers.; > >>>> input firstname lastname / > >>>> Address & / > >>>> @; > >>>> _infile_=reverse(_infile_); > >>>> input zip state city &; > >>>>cards; > >>>>John Doe > >>>>33 10 Av. > >>>>Uptown, Smalltown XX 12345 > >>>>;run; > >>>> > >>>>Regards, > >>>>Søren > >>>> > >>>>On Sun, 26 Dec 2010 11:07:11 -0500, Arthur Tabachneck > <art297@ROGERS.COM> > >>>>wrote: > >>>> > >>>>>The following was a question that was raised on the SAS discussion > >forum. > >>>>>You are confronted with data that has 3 lines per subject, but the > third > >>>>>line has variables that may contain embedded spaces, but there is only > >>one > >>>>>space between variables. > >>>>> > >>>>>The only suggestion I could think of was the one shown below. Is there > >>an > >>>>>easier way? > >>>>> > >>>>>data work.Address (drop=_:); > >>>>> infile cards; > >>>>> input FirstName $ LastName $ / > >>>>> Address $ 1 - 20 / > >>>>> _Third_Line & $80.; > >>>>> format City $10.; > >>>>> Zip=scan(_Third_Line,-1); > >>>>> State=scan(_Third_Line,-2); > >>>>> call scan(_Third_Line, -2, _position, _length); > >>>>> City=substr(_Third_Line,1,_position-1); > >>>>> cards; > >>>>>Lee Athnos > >>>>>1215 Raintree Circle > >>>>>New York NY 85044 > >>>>>Heidie Baker > >>>>>1751 Diehl Road > >>>>>Vienna VA 22124 > >>>>>; > >>>>> > >>>>>Thanks in advance, > >>>>>Art