Re: st: Simple and weird error

[Nick Cox <njcoxstata@gmail.com>]

This can't be literal code, as there are upper case letters and spaces
that would prevent this from running. Also, the local macro references
should all be of the form `macname' with different opening and closing
single quotation marks.

With a stylistic twist (one -quietly- instead of three) the code can
be corrected and rewritten

quietly foreach varname of varlist `r(varlist)' {
        replace `varname'  = log(`varname')
        gen lr`var' = `varname' - `varname'[_n-1]
        drop `varname'
}

But this wouldn't work either. `var' is the same each time round the
loop: either blank or whatever it was defined as before this code
segment. So, the code would fail second time round the loop.

Hence I give up. Your code as you report couldn't do more than
-generate- one variable, even with several corrections.

I suggest that you post your real code.

Nick

On Wed, Nov 21, 2012 at 10:54 PM, Zhangqunzi <zhangqunzi@googlemail.com> wrote:

> I am trying to calculate the log return lr
> according to the code below:
>
> For each varname of varlist 'r(varlist)'{
> Quiet replace 'varname' =log('varname')
>
> Quiet gen lr'var'='varname'[_n]-'varname'[_n-1]
>
> Quiet drop 'varname'
> }
>
> varname is price1,2,3 etc for each firm.
>
> In the end I found all the log return I get is positive which is clearly wrong for 5000 firms and 30years.
*
*   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/