Re: st: assymetric start and end dates in a panel dat set

[Nick Cox <njcoxstata@gmail.com>]

Please don't send attachments! This is explained in the FAQ.

By excel I presume you mean "MS Excel". Evidently you use it but many
Stata user-programmers emphatically do not. I don't want to look at
your spreadsheet files, thanks.

When I said "Look at -egen- for functions to count missings, or
non-missings, in observations or rows." I presumed that was enough of
a clue for you to identify the functions -rowmiss()- and
-rownonmiss()-.

After a -tsfill- you could do something like

egen nonmiss = rownonmiss(Greece Italy UK)

and then your analysis would be conditional

... if nonmiss == 3

No doubt you would have more countries and the number would not be 3.

Nick

On Thu, May 24, 2012 at 9:02 AM, loggyedr salvez
<loggyedy@googlemail.com> wrote:
> Thank you Nick.  Suppose that I correct the problem  and that I have
> observations every 4 weeks for all countries. Thus, I could apply the
> code you proposed-floor(/28)].  The problem is that  I have different
> start and end dates and different time periods across countries.
> For example
>              startdate-enddate
> Greece   02/11/06-09/10/09
> Italy        07/12/06-06/11/09
>  UK          09/11/06-06/11/09
>
> How should I handle this asymmetry?  I  Would be grateful to you if
> you could provide more information about which  functions  for –egen-
> to look at. Any proposed coding is also  welcomed as I am a new stata
> user and I have difficulty with coding. For example your proposed
> coding
>
> gen edate2 = floor(edate1/28)
> tsset identif2 edate2
>
> was extremely helpful. Should I upload an example of the excel file I
> have to take a look?
>
>
> Thank you for your time,
> stefanos
>
>
> On 5/24/12, loggyedr salvez <loggyedy@googlemail.com> wrote:
>> Thank you Nick. suppose that I correct the problem  and that I have 4
>> -weekly observations. Thus, I apply the code you proposed-floor(/28)]floo.
>> The problem still remains as I have different start and end dates and
>> different time periods across countries. I would be grateful to you if you
>> could give some more information as well as for the functions egen. I am a
>> new stata user and I have some difficulty  moving my work ahead.
>>
>> Thank you again
>>
>> On Wed, May 23, 2012 at 9:06 PM, Nick Cox <njcoxstata@gmail.com> wrote:
>>
>>> The answer to this question is surely based on the answer to your
>>> previous question. -tsset- in terms of a time variable that jumps
>>> every 28 days. Look at -egen- for functions to count missings, or
>>> non-missings, in observations or rows.
>>>
>>> Nick
>>>
>>> On Wed, May 23, 2012 at 8:59 PM, loggyedr salvez
>>> <loggyedy@googlemail.com> wrote:
>>>
>>> > I have a panel data set that refers to interest rates over time
>>> > accross countries.
>>> > I have observations every 4 weeks. For example
>>> >
>>> >  November 1, 2009
>>> > November 29, 2009
>>> > December 27, 2009
>>> > January 31, 2010
>>> >  November 1, 2009
>>> >  November 29, 2009
>>> >  December 27, 2009
>>> >  January 31, 2010
>>> >
>>> >
>>> > The problem is that the  start and the end date across countries is
>>> different.
>>> >
>>> > For example
>>> >               startdate-enddate
>>> > Greece   02/11/08-09/10/11
>>> > Italy        07/12/08-06/11/11
>>> >  UK          09/11/08-06/11/11
>>> >
>>> >
>>> > and furthermore as you can see the time interval is not the same for
>>> > all the countries (compare the time span of UK with that of Italy)
>>> > is there any solution to this problem? I want to regress interest
>>> > rates on other variables and if I do not know how to solve this
>>> > problem I might end up having wrong results. If I have to trim the
>>> > dates to obtain symmetry what would be the code for converting this
>>> > series of dates to numeric variable   ( I am referring to format time
>>> > %d, tsset identifier time and so forth)

*
*   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/