Hi,
Please see the example and the problem I am facing below. Any help in
getting around this issue will be much appreciated.
Thanks in advance,
Hugh
USING FILE
sysuse auto, clear
sort weight
drop if weight[_n]==weight[_n-1]
gen id_2 =_n
save auto1, replace
MASTER FILE
id_1 weight
1 1825
2 1825
3 2010
4 2017
5 2023
input id_1 weight
sort weight
nearmrg using auto1, nearvar(weight) genmatch(weight_merged) update
save master, replace
If you run this, you will find that the first two observations of the
MASTER file are matched with the same observation (i.e. weight=1830
from the USING file) as can be expected. Similarly, the last three
observations of the MASTER file are matched with the same observation
(i.e. weight=2020 from the USING file). This occurs because -nearmrg-
picks the closest match.
However, I require that all the five matches be unique (i.e. they
should not be repeated). In other words, for a given "weight" in the
MASTER file, I need the closest match from the USING file. The next
observation "weight" from the MASTER file (even if it equals the
previous "weight") must be directed to the next closest observation
from the USING file. The end result is that I should have five unique
matches.
*
