[数据处理]stata新人想了好久都想不出来。。求助

现在有模拟数据：

sample	x_1	x_2
1	1	50
1	1	41
1	3	38
1	4	60
1	10	65
1	3	71
1	2	21
1	2	30
1	1	40
1	6	66

找出sample里面top5（出现次数最多的前5个）x_1的x_2的平均值。如x_1=1,meanx_2=(50+41+40)/3
理想输出结果：

sample	top5x_1	meanx_2
1	1	(50+41+40)/3
1	2	(21+30)/2
1	3	(71+38)/2
1	4	60
1	6	66

各位这东西能用stata处理么TT？

gen x3=x1
collapse (count) x3 (mean) x2, by(sample x1)
gsort sample -x3 x1
by sample: keep if _n<=5

bys sample x1: egen meanx2=mean(x2)
bys sample: egen countx1=count(x1)
duplicates drop sample x1 countx1, force
bys sample: egen rankx1=rank(countx1), unique
keep if rankx1<=5
sort sample rankx1
keep sample x1 meanx2

another way:

tempvar c
egen double `c'=count(1), by(sample x_1)
egen double m=mean(x_2), by(sample x_1)
gsort sample -`c' x_1
keep if x_1!=x_1[_n-1]
bysort sample : keep if _n<=5