RE: st: RE: run through every values of a variable

[Nick Cox <n.j.cox@durham.ac.uk>]

Stata cares only about its rules, not your preferences or expectations. This is lesson #1! 

Nick 
n.j.cox@durham.ac.uk 

tashi lama
Subject: RE: st: RE: run through every values of a variable

I might add though....forvalues i=1/_N  should work at least in principle...because you are essentially sayin... forvalues i=1/14   if the last obs. is 14 and this is a perfectly sound syntax. In anycase, thanx..

----------------------------------------
> From: ltashi32@hotmail.com
> To: statalist@hsphsun2.harvard.edu
> Subject: RE: st: RE: run through every values of a variable
> Date: Thu, 3 May 2012 20:00:14 +0000
>
> count might just do the magic. Thank you. And yes, after second peek , I found whatever was in italic. I did realise that capture confirm wouldn't work in the following example since levelsof arranges the obs values as row and they aren't necessarily variables.
>
> levelsof broker_id, clean
>
> foreach item of local `r(varlist)' {
>
> if "`item'"=="555" {
>
> ..........
>
> }
>
> }
>
>
>
> might work...
>
> Thanks again
>
>
>
> ----------------------------------------
> > From: n.j.cox@durham.ac.uk
> > To: statalist@hsphsun2.harvard.edu
> > Date: Thu, 3 May 2012 20:25:19 +0100
> > Subject: st: RE: run through every values of a variable
> >
> > Your -confirm- statement is illegal syntax, and putting -capture- in front will just eat the syntax error, so that is not the way to go.
> >
> > If you want to check whether any value is 555 then you can just go (e.g.)
> >
> > assert broker_id != 555
> >
> > count if broker_id == 555
> >
> > These are some of the ways are finding whether any value is 555. The first will be contradicted if any value is 555. The second will return a positive count in the same circumstance.
> >
> > Using a loop is _very_ inefficient here. Your loop is illegal because -forvalues- doesn't understand _N as such; you need to evaluate it as `=_N'.
> >
> > Your references are not getting any more precise! "stata resource and support page" does not remind me where I wrote that, as (let me put this in a David Hoaglin-like understatement) I have written various pieces on Stata here and there, including various FAQs.
> >
> > But no matter: whatever is a placefiller indicating some generic command. There is no official Stata command -whatever-. If you look closely at what was written, "whatever" should be in italic or indicated by <whatever>. If not, let me know where you found it and we will fix it.
> >
> > Nick
> > n.j.cox@durham.ac.uk
> >
> > tashi lama
> >
> > I have a variable broker_id in my dataset as follows...
> >
> > broker_id
> >
> > 6429
> > 131
> > 244
> >
> > and I have to see if "555" matches any of the values in a broker_id. Now, I think I could use some technique like
> >
> > levelsof broker_id, clean
> > capture confirm var "555"
> >
> > but I am wondering if there is a st. forward way to do this using forvalues loop like this
> >
> > 1. When I did forvalues i=1/_N {
> >
> > if broker_id[`i']=="555" {
> >
> > ........
> >
> > I get invalid syntax. I am not so sure... Could someone explain?
> >
> > 2. I see that Nick has given a way to do this in stata resource and support page
> >
> > sysuse auto
> >
> > foreach i in 0 1 {
> > whatever if foreign == `i'
> >
> > }
> >
> > I tried to run this and got unrecognized command whatever in return. Is it because whatever is an obsolete command? Any idea....

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/