That's summing rowwise. I see no signal that that is what is wanted.
Since Stata 9 or thereabouts, -egen-'s -rowsum()- function has been
undocumented in favour of -rowtotal()-, but it still works well.
Nick
On 8 Feb 2013, at 17:14, "Dudekula, Anwar" <dudekulaan@upmc.edu> wrote:
Hi Mike
Can you use egen(sumperyear): rowsum(var1-var5)
Thank you,
Yours Sincerely,
Anwar
________________________________________
From: owner-statalist@hsphsun2.harvard.edu [owner-
statalist@hsphsun2.harvard.edu] on behalf of Nick Cox
[njcoxstata@gmail.com]
Sent: Friday, February 08, 2013 9:36 AM
To: statalist@hsphsun2.harvard.edu
Subject: Re: st: FW: forvalues question...
Yes, that's better than my solution. I got hung up on how to do it
with -forvalues-, when a better answer is to do it directly.
Nick
On Fri, Feb 8, 2013 at 2:27 PM, Scott Merryman <scott.merryman@gmail.com
> wrote:
On Fri, Feb 8, 2013 at 8:04 AM, Mahometa, Michael J
<michael.mahometa@ssc.utexas.edu> wrote:
All,
I've got a dataset that has multiple rows for one year, with
several sequential variables. And, I'd like to sum what happens
over the year (using something like collapse). I have 20 variables
and the data looks like:
Year var1 var2 var3 var4 var5 ....
1999 2 5 6 1 4
1999 4 1 5 2 5
1999 1 3 5 6 7
2000 6 5 1 3 8
2000 1 1 2 3 6
...
So the current "brute force" method is:
collapse (sum) var1_s=var1 var2_s=var2 ///
var3_s=var3 var4_s=var4, by(year)
But this seems in-elegant.
How about:
collapse (sum) var*, by(year)
rename var* =_s
