Thanks, Martin and Michael, both for your hints and for the exegetic
exercise on my words, which I found amusing.
-rolling- seems to be doing the job I want, especially if combined
with "[r]recursive". The problem I see there is that I can perform
only one command within the "frame" of this rolling window, not a
series of commands (like postestimation commands), as I was thinking
of when specifying it as a loop.
That's why I agree with Michael's points, and I find your solution,
Martin, to use a loop beforehand to fill a -local- with the desired
end dates very promising.
@ Martin: I did not really care about what kind of list, I was indeed
just trying to explain my goal in pseudo-code. So am I allowed to
rephrase my suggestion before, inventing a new type of list, call it
"timelist"? That would have been my dream (and to all German readers:
please note the "pseudo" in "pseudo-code"):
*** begin pseudo-code
foreach t of timelist 1992q1/2000q4 {
reg y L.y L.x if tin(1960q1, `t')
}
*** end pseudo-code
Best, and thanks again! AC
On Wed, Dec 3, 2008 at 4:34 PM, Martin Weiss <[email protected]> wrote:
> BTW, Augusto did not even pretend that his was a -numlist-; if he had
> thought it was one, he would have said something like "...of numlist
> blabla". His example looks like "any_list" in the -h foreach- parlance. He
> could of course use a loop beforehand to fill a -local- with the desired end
> dates and use that in the -foreach- ...
>
> HTH
> Martin
> _______________________
> ----- Original Message ----- From: "Michael Hanson" <[email protected]>
> To: <[email protected]>
> Sent: Wednesday, December 03, 2008 10:29 PM
> Subject: Re: st: Re: Running a foreach loop over a sequence of dates
>
>
>> Ah, so true -- but by labeling it as "pseudo-code", Augusto indicated that
>> he already knew that. Nonetheless, looping over dates would be *extremely*
>> high on my list of wishes & grumbles! Perhaps in Stata 11.... or even 10.2.
>> (Since, in the end, dates are just formatted numbers, this seems as if it
>> should be possible in theory.) There are applications in which something
>> much more light-weight than - rolling- would be appropriate. Although I
>> agree that the example below could likely be accomplished with -rolling-.
>>
>> -- Mike
>>
>> On Dec 3, 2008, at 4:05 PM, Martin Weiss wrote:
>>
>>> You cannot simply abbreviate your list of end dates like a - numlist-
>>> with a slash and hope for Stata`s acquiescence... I would recommend you
>>> look at -h rolling-
>>>
>>> HTH
>>> Martin
>>> _______________________
>>> ----- Original Message ----- From: "Augusto Cadenas"
>>> <[email protected]>
>>> To: <[email protected]>
>>> Sent: Wednesday, December 03, 2008 10:02 PM
>>> Subject: st: Running a foreach loop over a sequence of dates
>>>
>>>
>>>> Hello,
>>>>
>>>> I am trying to run a series of commands in a time series setting, and
>>>> interested in using variable time intervals (variable time windows, or
>>>> variable start/end dates). So ideally, I would like to loop with
>>>> -foreach- over a series of dates, but I cannot find the way of doing
>>>> this. To give you an idea of the kind of problem, here it is in
>>>> pseudo-code (not working):
>>>>
>>>> *** begin pseudo-code
>>>>
>>>> foreach t in 1992q1/2000q4 {
>>>> reg y L.y L.x if tin(1960q1, `t')
>>>> reg y L.y L.x if tin(`=`t'-20', `t')
>>>> }
>>>>
>>>> *** end pseudo-code
>>>>
>>>> I am sorry if this is a trivial question, but I really could not find
>>>> a way of performing this. Thanks for any hints.
>>>>
>>>> AC
>>>> *
>>>> * 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/
