-strparse- was the precursor of official Stata's
-split-. From what I can remember the syntax
would be identical here.
Nick
[email protected]
[email protected] replied to Barbara Hofmann <[email protected]>
> > i want to read in a ascii-dataset of a structure like this:
> >
> > 104525300010 9064496 022 999929912
> > 204525300010 9064496 1:19 59 22 2:4 0 96 3:8 0 92 4:0 33 67
> > 204525300010 9064496 6:30 50 20 7:78 5 17
> > 104525300010 9081400 023 23131122
> > 104525300020 7 332321332 06 00 05 2264
> > 104525300020 8 221211121 06 00 09 1889
> > 104525300020 8 222222222 06 00 08 4705
> > 104525300020 8 231211222 06 00 06 4585
> >
> >
> > "2" in the first column means that this line belongs to the line
> > before. column 31-39 include replies to 9 different questions. if
> > it is a "9" this means, that in the following line there is a more
> > specific answer to this question which is stored like this:
> > 7:78 5 17 means "78% replied to question 7 with a "1", 5% replied
> > to question 7 with a "2" and 17% replied to question 7 with a "3""
> >
> > Then ":" signalizes that the answers to the question of the number
> > before ":" will follow. How can i tell Stata this meaning of ":"?
> > Maybe also a problem could be that the answers can be like "10 90
> > 0" as well as like "15 15 70" or "5 5 90"?
> >
> As a start, the commands below, I believe, pull out the
> relevant information.
> It uses Michael Blasnik and Nick Cox’s –strparse- command.
>
> insheet using "C:\Documents and
> Settings\scott.merryman\Desktop\foo.txt", clear
> gen id = substr(v1, 1,1)
>
> forv i = 31/39 {
> gen question`i' = substr(v1, `i', 1) if id == "1"
> }
>
> gen answer = substr(v1, 23,.) if id == "2"
>
> strparse answer, g(b)
>
> forv i = 1/12 {
> strparse b`i' , gen(detail`i') parse(:)
> }
>
> drop b*
>
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
