Thomas,
I hope I understood you problem correctly. Here is my solution to it:
// create variable with all entries set to missing
gen first=.
//check each entry within group (sorted by time) whether it is
//non-zero and non-missing and whether it is the first one which
//fulfills this condition, if so record the _n for it to variable first
//if not so record the entry in first from the preceding observation
//in it.
bys group (time): replace first=cond(expression>0 & expression<. &
first[_n-1]==.,_n,first[_n-1])
//drop all obs with first == missing
drop if first==.
hth
Sebastian
On 2/4/07, Thomas Speidel <thomas@tmbx.com> wrote:

I am trying to tag the first non-zero, non-missing observation within a
variable by group. The goal is to drop anything BEFORE the expression
is measured within a group. For example:
group time expression
1 1 0
1 2 0
1 3 3.2
1 4 4.1
1 5 0
1 6 2.5
2 1 0
2 2 .
2 3 0
2 4 3.7
2 5 4.0
..
My goal is to obtain this:
group time expression
1 3 3.2
1 4 4.1
1 5 0
1 6 2.5
2 4 3.7
2 5 4.0
..
I would like to avoid reshaping the dataset (where the useful "egen
rfirst" would help). I have tried using various subscripting
expressions, but none produced the end result.
bysort group: gen firstexp=1 if expression[_n] > expression[n-1]
bysort group: gen cum=sum(firstexp)
gen tag=firstexp-cum
gen first=tag==0
The above commands gets me one step closer, but it still doesn't get me
there...
Thanks,
Thomas Speidel
*
* 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/

*
* 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/