Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | Michael McCulloch <mm@pinestreetfoundation.org> |
To | statalist@hsphsun2.harvard.edu |
Subject | Re: st: editing string variables to remove letters and keep only numbers |
Date | Tue, 18 Jun 2013 10:34:05 -0700 |
Thanks Nick and Rich, These comments helped me get started on understanding forval. And in Nick's interpretive instructions, I discovered my error: If I run this entire section of code in one do-file command, it is successful: gen length = length(id) su length, meanonly local lmax = r(max) gen numstr = "" gen letterstr = "" forval i = 1/`lmax'{ replace numstr = numstr + substr(id,`i',1) if inrange(real(substr(id,`i',1)),0,9) replace letterstr = letterstr + substr(id,`i',1) if !inrange(real(substr(id,`i',1)),0,9) } However, when i was running the first 5 lines and then the forval separately, the syntax error message occurred. Why is that? Best wishes, Michael McCulloch, LAc MPH PhD -- Pine Street Foundation, since 1989 124 Pine Street | San Anselmo | California | 94960-2674 P: (415) 407-1357 | F: (206) 338-2391 | http://www.PineStreetFoundation.org On Jun 18, 2013, at 10:20 AM, Nick Cox wrote: > Rich is right. The example code was meant to show how simple the idea > was; it seems that was a failure. > > Starting again, and correcting a typo of mine. > > #1 What's the length of the longest string? > > #1.1 What is the length of each string value? > gen length = length(strvar) > > #1.2 Summarize that > su length, meanonly > > #1.3 Put the maximum in a local macro > local lmax = r(max) > > #2 Initialise variables to hold numeric and non-numeric characters: > gen numstr = "" > gen letterstr = "" > > > #3 loop over characters from first to last > forval i = 1/`lmax' { > > #3.1 one command line only > # if i th character is between 0 and 9, it's a numeric character, so > add it to numeric character variable > > replace numstr = numstr + substr(strvar, `i', 1) if > inrange(real(substr(strvar, `i', 1)), 0, 9) > > #3.2 one command line only > #otherwise add it to non-numeric character variable > > replace letterstr = letterstr + substr(strvar, `i', 1) if > !inrange(real(substr(strvar, `i', 1)), 0, 9) > > } > > Nick > njcoxstata@gmail.com > > > On 18 June 2013 17:03, Richard Goldstein <richgold@ix.netcom.com> wrote: > >> looks to me like you are putting what should be on one line onto two >> lines without the "///" continuation; note that in the code below, the >> loop should only have two lines, each starting with "replace" > > On 6/18/13 12:00 PM, Michael McCulloch wrote: > >>> Nick, as it's NOW written (with your correction included), I still receive an "invalid syntax" error. >>> This is my first exercise in learning -forval-, so even after reviewing the -help- I'm not sure how to troubleshoot further. >>> >>> forval i = 1/`lmax' { >>> replace numstr = numstr + substr(id, `i', 1) if >>> inrange(real(substr(id, `i', 1)), 0, 9) >>> replace letterstr = letterstr + substr(id, `i', 1) if >>> !inrange(real(substr(id, `i', 1)), 0, 9) >>> } >>> > * > * For searches and help try: > * http://www.stata.com/help.cgi?search > * http://www.stata.com/support/faqs/resources/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/faqs/resources/statalist-faq/ * http://www.ats.ucla.edu/stat/stata/