If the identifier is missing on lines 2 to whatever
of each block, you would need to fill in the missings
_before_ you did this. Otherwise your dataset will
be messed up.
The fill in is just
replace identifier = identifier[_n-1] if mi(identifier)
Nick
[email protected]
Ada Ma
> You can use -reshape- as you say, but before you do it you need to
> generate a variable which shows which record the obs correspond to
> amongst all the obs you have for each person, something like the count
> variable in here:
>
> identifier dob request count
> ss12345 10 04 68 MA4 1
> A2 2
> ST69979 3
> verygood 4
> ss12346 21 3 52 . 1
> ss12347 14 8 71 MA2 1
>
> You said you don't care about the particular order of the request
> variable so you can write:
>
> bys identifier: gen count=_n
> reshape wide request, i(identifier) j(count)
> On 5/15/06, Eamonn O'Brien <Eamonn.O'[email protected]> wrote:
> > I have a data set that I would like to 'reshape' to a
> record per row.
> > All variables except one, the 'request' variable are on a
> single row per
> > identifier. The request variable can be any number of rows
> and contains
> > both string and numeric data, with the codes I am interested in
> > appearing in no particular order in the column. I would like to
> > reshape this variable so it's on one row with the identifier. I have
> > explored -xpose- and -sxpose- and reshape but cannot achieve this.
> > (Their are more than 3 variables in the data set).
> >
> > identifier dob request
> > ss12345 10 04 68 MA4
> > A2
> > ST69979
> > verygood
> > ss12346 21 3 52 .
> > ss12347 14 8 71 MA2
> >
*
* 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/
