Re: st: grouping variable

[Nick Cox <njcoxstata@gmail.com>]

For quick inspection you could always

bysort id (admission_date) : gen first = _n == 1
edit if first

Of course that misses much of the detail but it's a quick way of
getting one observation for each id

On Thu, Jan 12, 2012 at 11:45 AM, Lars Folkestad
<lfolkestad@health.sdu.dk> wrote:
> Thank you Nick for your answer.
>
> My id is the social security number of each individual. For making the
> data easier to read here in the initial phase of my work
> I would like to have the data like this:
>
> ID sex admission date (admission1) hospital (admission1) department
> (admission1) ... Department (admissionN)
>
> Instead of the way data is now:
> Id sex admission date hospital department
> 1   1   DDMMYY          1       1
> 1   1   DDMMYY          1       2
>
> And so forth.
>
> Your code (as always) did the trick.
>
> lars
>
>
>
> Den 12/01/12 12.32 skrev "Nick Cox" <njcoxstata@gmail.com>:
>
>>There is no rule that -i()- must specify a single variable. In your
>>case however you probably want a new sequence variable
>>
>>bysort id (date_admission) : gen seq = _n
>>
>>and then to -reshape- using -i(id seq)- (not -i(id)-). Getting
>>admissions on the same day in the right order sounds tricky unless you
>>also have a time-of-day variable.
>>
>>That said, this kind of -reshape- usually makes later analysis more
>>difficult, so exactly why you think it will help you is an open
>>question.
>>
>>Nick
>>
>>On Thu, Jan 12, 2012 at 11:19 AM, Lars Folkestad
>><lfolkestad@health.sdu.dk> wrote:
>>
>>> After searching the web i will have to ask you - co-listers.
>>>
>>> I have a dataset of patients and admissions.
>>>
>>> Id sex hospital ward date_admission date_discharge
>>>
>>>
>>> Data is in long format and i would like it to be reshaped to wide
>>>format.
>>> Some participants have up to 150 different admissions.
>>>
>>> My problem is that i dont have a unique grouping variable available
>>>(some
>>> patients have been admitted to the same wards twice or more on the same
>>> day)
>>>
>>> I would like to do the following
>>>
>>> Sort by id
>>> Genereate a grouping variable 1-_n for each id
>>> Reshape the lot to wide using i(id) j(groupvar)
>>>
>>> But i cannot se how to to this.
>>>
>>> Any other ways do reshape?

*
*   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/