Duplicates drop

[Naveen Adhikari]

Hello all, I am trying to drop the duplicate id. For example in below i have computed the total of expenses for each categories for a household using egen var= total(var), by(WWWHH). Now i want the total for each variables corrosponding to each id. However, as can be seen in data, for id 1, the total of grains is appeared in first few rows while pules expenditure appears thereafter and so on. If i use the command duplicates drop then my pulses_exp milk_exp will all be missing. How i can total for example like
id grains_exp pulse_exp
1 2544 160
and so on.
,

[Carlo Lazzaro]

Naveen:
you may want to try:

set obs 10
g id=1
g grain_exp=10 in 1/5
g pulse_exp=5 in 6/10
collapse (mean) grain_exp pulse_exp, by( id)///advice: save your dataset in a new file before -collapse-
su grain_exp pulse_exp
egen total=rowtotal( grain_exp pulse_exp)

[Oded Mcdossi]

What you need here is to collapse your data by id:

collapse (max) grains_exp pulses_exp milk_exp, by(WWWHH)

[Joseph Coveney]

Well, here's one way

tempfile tmpfil0

// Primes the pump and allows trapping of all-missings
preserve
contract WWWHH, freq(discard)
quietly save `tmpfil0'
restore

foreach var of varlist *_exp {
    preserve
    contract WWWHH `var', freq(discard)
    keep if !mi(`var')
    merge 1:1 WWWHH using `tmpfil0', assert(match) nogenerate noreport // Traps duplicates, also
    quietly save `tmpfil0', replace
    restore
}

use `tmpfil0', clear
drop discard

[Naveen Adhikari]

Thank you all, it really helped me.