This leaves open the question of what
to do if these dates are just strings.
One solution looks like this:
gen date2 = ym(year(date("1 " + bclmeddt, "dmy", 2040)),
month(date("1 " + bclmeddt, "dmy", 2040)))
Is there a more direct way to do it?
Nick
[email protected]
> Joao Pedro W. de Azevedo
>
> > I'm trying to generate a new dataset and I would sincerely
> > appreciate
> > suggestions or how to approach the problem.
> > I have records on individuals start and end date on a
> > program participation
> > on a long format (table 1) and I'm trying to reshape it to
> > a wide format in
> > order to generate a table similar to TABLE 2. Table 2 takes
> > the value of one
> > for the monthly intervals encompassed between the start
> and end date
> > available on tabe 1 and zero for the remaining periods.
> >
>
> No doubt -reshape- could be used here but a different
> approach springs to mind. The reason is important: even if
> you got to a wide structure in which each month is a variable,
> what would you use to tabulate those variables? (There is
> an answer to this!)
>
> I assume that underneath the format these are really
> Stata monthly dates.
>
> . l
>
> +-----------------------------+
> | serial bclmsdt bclmeddt |
> |-----------------------------|
> 1. | 2 May 99 Jul 99 |
> 2. | 20 Jul 99 Nov 99 |
> 3. | 23 Aug 99 Feb 00 |
> 4. | 26 Oct 99 Dec 99 |
> 5. | 30 May 99 Jun 99 |
> |-----------------------------|
> 6. | 30 Dec 99 Feb 00 |
> +-----------------------------+
>
> . preserve
> . gen diff = bclmeddt - bclmsdt
> . expand diff
> . bysort serial : replace bclmsdt = bclmsdt[_n-1] + 1 if _n > 1
> . tab serial bclmsdt
> . save expanded
> . restore
*
* For searches and help try:
* http://www.stata.com/support/faqs/res/findit.html
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
