Although you report success, there seems to be a difficulty with your
approach.
By construction, values with the same value of -grp- all have the same
value of -year-. Conversely, observations in the same group of -grp-
cannot span two calendar years.
Focus on the qualifier
... if month>m(`a'm4) & month <= m(`b'm4) & grp==`d'
Here a is the year corresponding to d and b is a + 1; however, it
follows that no observation can be in year b and also in group d.
Seemingly, this problem was not fatal.
With similar problems in the past, I've found the best approach to
subdivide according to the non-calendar years that are the time units.
For example, there is an -egen- function in -egenmore- with that intent.
The principles behind it may be relevant to your problem.
Nick
[email protected]
Richard Harvey
I managed to arrive at the solution, i made two modifications to the
code. I created a indicator variable and then dropped the observations
from outside the loop also i changed the condition statement.
egen grp=group(id year)
gen ind=.
set trace on
levelsof grp, local(m)
foreach d of local m {
summ year if grp==`d'
local a=`r(mean)'
local b=`a'+1
di `a' " " `b'
replace ind=1 if month>m(`a'm4) & month <= m(`b'm4) & grp==`d'
}
drop if ind~=1
and that gives me exactly what i want..
id year month var1 grp ind
1 1980 1980m4 7 1 1
1 1980 1980m5 14 1 1
1 1980 1980m6 3 1 1
1 1980 1980m7 16 1 1
1 1980 1980m8 18 1 1
1 1980 1980m9 4 1 1
1 1980 1980m10 2 1 1
1 1980 1980m11 0 1 1
1 1980 1980m12 13 1 1
1 1980 1981m1 20 1 1
1 1980 1981m2 11 1 1
1 1980 1981m3 13 1 1
1 1980 1981m4 21 1 1
1 1981 1981m4 21 2 1
1 1981 1981m5 12 2 1
1 1981 1981m6 25 2 1
1 1981 1981m7 7 2 1
1 1981 1981m8 4 2 1
1 1981 1981m9 16 2 1
1 1981 1981m10 4 2 1
1 1981 1981m11 19 2 1
1 1981 1981m12 16 2 1
1 1981 1982m1 17 2 1
1 1981 1982m2 3 2 1
1 1981 1982m3 17 2 1
1 1981 1982m4 17 2 1
2 1980 1980m4 12 3 1
2 1980 1980m5 1 3 1
2 1980 1980m6 8 3 1
2 1980 1980m7 13 3 1
2 1980 1980m8 2 3 1
2 1980 1980m9 8 3 1
2 1980 1980m10 17 3 1
2 1980 1980m11 14 3 1
2 1980 1980m12 35 3 1
2 1980 1981m1 7 3 1
2 1980 1981m2 4 3 1
2 1980 1981m3 21 3 1
2 1980 1981m4 11 3 1
2 1981 1981m4 11 4 1
2 1981 1981m5 1 4 1
2 1981 1981m6 17 4 1
2 1981 1981m7 22 4 1
2 1981 1981m8 8 4 1
2 1981 1981m9 31 4 1
2 1981 1981m10 16 4 1
2 1981 1981m11 20 4 1
2 1981 1981m12 4 4 1
2 1981 1982m1 13 4 1
2 1981 1982m2 9 4 1
2 1981 1982m3 15 4 1
2 1981 1982m4 24 4 1
*
* For searches and help try:
* http://www.stata.com/help.cgi?search
* http://www.stata.com/support/statalist/faq
* http://www.ats.ucla.edu/stat/stata/
