Martin, it applies not only to -describe-, it is all over:
sysuse auto, clear
rename trunk income
regress price inc*
regress price in*
And in this case the message "varlist not allowed" can knockout anyone
who wanted to use it in according to the syntax:
regress depvar indepvarlist
It also applies to reserved words -if- and -using-, but not to -byte-.
(Tried "ifo", "using_method", and "bytevar" as variable names in the
code above and abbreviated to the length of the reserved word).
In general I do want the above regression to work, and when parsing a
command for parts of syntax, "in" should be different from "in*". But
may be there is something else there. Since -syntax- is built-in , we
may never know, unless someone from Stata, corp steps in.
A related 'bug' (?) would be the following:
generate byte long float double z=1
is this a valid command? if yes, what is the resulting type of z? why?
Best wishes,
Sergiy Radyakin
On Wed, Oct 29, 2008 at 5:34 PM, Martin Weiss <[email protected]> wrote:
> Ok, thanks for that! I thought that line 7 of describe.ado did the same as
> -varlist- in -syntax- but your example has convinced me that it does not.
> What is your take on the bug issue, then?
>
>
> HTH
> Martin
>
>
> -----Original Message-----
> From: [email protected]
> [mailto:[email protected]] On Behalf Of Sergiy Radyakin
> Sent: Wednesday, October 29, 2008 10:26 PM
> To: [email protected]
> Subject: Re: st: RE: minor bug in -describe-?
>
> Hi Martin,
>
> if one writes -varlist- variables must already exist in memory. But
> you sometimes want:
>
> clear
> describe somevars using somefile.dta
>
> to work.
>
> Your modified version will work ok, but replace it with varlist and
> see what happens.
>
> Best regards,
> Sergiy Radyakin
>
>
>
>
> On Wed, Oct 29, 2008 at 4:46 PM, Martin Weiss <[email protected]> wrote:
>> Curiously, describe.ado takes -anything- in its -syntax- statement. If you
>> copy it to your very own describe2.ado in your personal directory and
>> restate that as -anything(everything)- then you can pass it "in*" and it
>> works as expected. I think that could be called a bug. Whether minor or
> not,
>> I do not want to decide... -help describe- indicates that it takes a
>> -varlist- so why would one want -anything- and not -varlist- in the
> -syntax-
>> statement?
>>
>>
>> HTH
>> Martin
>>
>>
>> -----Original Message-----
>> From: [email protected]
>> [mailto:[email protected]] On Behalf Of Jeph Herrin
>> Sent: Wednesday, October 29, 2008 6:53 PM
>> To: [email protected]
>> Subject: st: minor bug in -describe-?
>>
>>
>>
>> -in- is a reserved word in Stata, but it's also a very
>> common prefix for variable names. So I often run into
>> this:
>>
>> . d in*
>> varlist not allowed
>> r(101);
>>
>>
>> Obviously, -describe- takes a varlist, but it won't take
>> a varlist that looks like a reserved word. A bug, right?
>> Or is there a reason for this?
>>
>> thanks,
>> Jeph
>>
>>
>>
>>
>>
>> *
>> * 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/
>>
>> *
>> * 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/
>>
> *
> * 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/
>
> *
> * 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/
>
*
* 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/
