This is a bit too much like a puzzle in which you
have to guess the puzzle as well as the solution.
The rules seem to be
1. Ignore my typos. (The data for dif = 2, sector = 2
seem out of line.)
2. Treat duplicates on -dif- and -sector- as if
they occurred just once.
3. Then calculate cumulative sums over -sector-.
On those guesses, I guess in turn
bysort sector dif : gen CAAR = _n == 1
by sector : replace CAAR = sum(CAAR * AAR)
separate CAAR, by(sector) gen(CAAR_)
In general, -by:- is indispensable.
Nick
[email protected]
mai mai
> I am trying to get a cumulative sum of AAR below by sector group and
> diff group. So need to create four variables: the first (CAAR_1)
> containing the cumulative sum by dif if the firms were in sector 1,
> the second (CAAR_2) contains the cumulative sum by dif if the firms
> were in sector 2, and so on: I need it to do something like this:
>
> dif sector AAR CAAR_1 CAAR_2 ....CAAR_4
> 1 1 2.2 2.2 .
> 1 1 2.2 2.2 .
> 1 2 3.2 . 3.2
> 1 2 3.2 . 3.2
> 1 2 3.2 . 3.2
> 1 3 1.0 . .
> 1 3 1.0 . .
> 2 1 6.6 8.8 .
> 2 1 6.6 8.8 .
> 2 2 3.0 . 5.2
> 2 2 6.6 . 5.2
> 2 2 6.6 . 5.2
> 2 3 6.6 . .
> 2 3 6.6 . .
> 3 1 1.4 10.2 .
> 3 1 1.4 10.2 .
> 3 2 1.1 . 6.3
> 3 2 1.1 . 6.3
> 3 2 1.1 . 6.3
> 3 3 1.4 . .
> 3 3 1.4 . .
>
>
> I cannot do by line because because the groups vary in number of
> sub_occupants. I need to do it by group, can you help me?
*
* 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/
