Tak Wai,
It is not clear what you mean when you speak of observations and
variables. Does each v* variable have the same value in all
observations? When you say "v5 is missing", does that mean that it
has missing values in all observations? How exactly does your data
look like? What is the logic behind what you are trying to
accomplish?
If you want to count the number of missing values across all v*
variables for each observation, you can use -egen-.
. egen s = rowmiss(v1-v10)
Friedrich
--- Tak-wai Chau <[email protected]> wrote:
> Hi, Friedrich and other statalisters,
>
> Thanks for Friedrich's reply.
>
> Sorry that I made a mistake in my code. I mean "gen s=1" instead of
>
> "s=1". The missing values appear at different variables for
> different
> observations, so s would take different values for different
> observations. Also, what I want is that, for example, if all Vi are
> not
> missing except V5, then the resulting Wi would be that W1-W4 would
> be the same as V1-V4, while W5-W9 would carry values from V6-V10.
>
> Thanks a lot!
>
> Tak Wai
>
> Friedrich Huebler wrote:
> > Tak Wai,
> >
> > The loop below copies all non-missing observations from variables
> v1
> > to v10 to variables w1 to w10. The macro s counts the number of
> > non-missing observations in variables v1 to v10.
> >
> > local s = 0
> > forvalues i = 1/10 {
> > replace w`i' = v`i' if v`i'<.
> > count if v`i'<.
> > local s = `s' + r(N)
> > di `s'
> > }
> >
> > The statement "s=1" is not valid Stata, you probably mean
> "generate
> > s=1". The command "local s1=s" assigns the first observation in
> > variable s to the local s1. The command "replace s=s+1 if v`i'<."
> > adds 1 to all observations in variable s if the first observation
> in
> > variable v`i' is missing.
> >
> > The command "replace w`s1'=v`i' if v`i'<." indicates that you may
> be
> > thinking in terms of a loop over observations, not variables.
> Such a
> > loop is neither efficient nor necessary but if you want it you
> need
> > different code from the one above.
> >
> > Friedrich
> >
> > --- Tak-wai Chau <[email protected]> wrote:
> >> Hi,
> >>
> >> I would like to ask something about using macro in loops. I have
> >> tried to find it in the archive and FAQ for some time but I
> don't
> >> get it.
> >>
> >> Sorry if I have overlooked some obvious things.
> >>
> >> I would like to update certain variables according to the value
> >> taking by a particular variable, and this value differs by
> >> observations. Then what I should do?
> >>
> >> For example, if I want to transfer value of non-missing
> variables
> >> from V1-V10 to W1-W10. But I don't want to variables carrying
> >> missing value between. I use s to count the number of
> non-missing
> >> values I have covered at that point.
> >>
> >> The following is what I think of but doesn't work.
> >>
> >> s=1
> >> forvalues i=1/10 {
> >> local s1=s
> >> replace w`s1'=v`i' if v`i'<.
> >> replace s=s+1 if v`i'<.
> >> }
> >>
> >> I know the problem is at the local statement.
> >>
> >> What should I do?
> >>
> >> Thank you very much!
> >>
> >> Tak Wai Chau
____________________________________________________________________________________
No need to miss a message. Get email on-the-go
with Yahoo! Mail for Mobile. Get started.
http://mobile.yahoo.com/mail
*
* 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/
