Gayatri Koolwal posted:
> Hello,
>
> I have a dataset which includes three variables as follows:
>
>
> var 1 2 3
>
> hhd1 a n o
> hhd1 b o .
>
> hhd2 c p r
> hhd2 d q .
> hhd2 e r .
>
> hhd3 f s o
> hhd3 g o .
> .
> .
> .
> .
>
>
> that is, variable 2 and 3 share common values (like an id variable), with
> some missing values for variable 3. i would like to create a fourth variable
> which gives the value for variable 1 (originally in the same row as the
> observation variable 2) in the same row as the observation for variable 3:
>
>
>
> var 1 2 3 4
>
> hhd1 a n o b
> hhd1 b o . .
>
> hhd2 c p r e
> hhd2 d q . .
> hhd2 e r . .
>
> hhd3 f s o g
> hhd3 g o . .
>
>
> but since the common 'id' variable may be the same for different individuals
> across households, i also wanted to correct for that when matching columns 3
> and 2 (so that 'g' does not get matched to household 1 instead of household
> 3). is there a way to join these two functions together in Stata?
>
> Thank you very much,
>
> Gayatri Koolwal
>
>
>
--------------------------------------------------------------------------------
How about something like in the following do-file?
Joseph Coveney
--------------------------------------------------------------------------------
* next few lines are just inputing the dataset
clear
input str4 not_for_use str1 obs_no str1 id1 str1 id2
hhd1 a n o
hhd1 b o
hhd2 c p r
hhd2 d q
hhd2 e r
hhd3 f s o
hhd3 g o
end
* the action starts here
sort obs_no
generate byte id3 = id2!=""
replace id3 = sum(id3)
bysort id3: replace id2 = obs_no[_N]
by id3: replace id2 = "" if _n > 1
exit
--------------------------------------------------------------------------------
*
* 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/
