g******7 发帖数: 1433 | 1 I have a dataset, and I want to randomly split it into two datasets.
For example,
Obs Policy #
1 67
2 67
3 67
4 78
5 78
....
10000000 77821178
10000001 77821178
10000002 77821178
I want all the same (unique) policy# in one splitted dataset (like all the
67 in first dataset), not in both of splitted datasets(like obs 1 in
dataset1, obs 2 in dataset2), how would I do that in SAS?
Sorry that i cant type chinese,
Thanks! |
a****g 发帖数: 8131 | 2 proc sql/freq to find unique policy numbers
then proc surveyselect
【在 g******7 的大作中提到】 : I have a dataset, and I want to randomly split it into two datasets. : For example, : Obs Policy # : 1 67 : 2 67 : 3 67 : 4 78 : 5 78 : .... : 10000000 77821178
|
g******7 发帖数: 1433 | 3 could you write me the sample code on how to do this?
never write this kind of codes..
I have a large amount(millions) of unique policy #, will this gonna work?
baozi xie!
【在 a****g 的大作中提到】 : proc sql/freq to find unique policy numbers : then proc surveyselect
|
a****g 发帖数: 8131 | 4 it is somethinglike the following
proc sql;
create table a as
select distinct policy
from data;
quit;
proc surveyselect data=data method=random out=out1;
where policy in (proc sql; select policy from a;)
....
then
proc sql;
create data out2
as
select * from
data
where policy not in (select policy from out1);
quit;
something like this. haven't used it for a long time don't remember some
minute details
【在 g******7 的大作中提到】 : could you write me the sample code on how to do this? : never write this kind of codes.. : I have a large amount(millions) of unique policy #, will this gonna work? : baozi xie!
|
g******7 发帖数: 1433 | 5 不是很清楚第二步
proc surveyselect data=data method=random out=out1;
where policy in (proc sql; select policy from a;)
这里的method是random? 还有syntax具体是怎么样的?
【在 a****g 的大作中提到】 : it is somethinglike the following : proc sql; : create table a as : select distinct policy : from data; : quit; : proc surveyselect data=data method=random out=out1; : where policy in (proc sql; select policy from a;) : .... : then
|
D******n 发帖数: 2836 | 6 proc surveyselect;cluster;
【在 g******7 的大作中提到】 : I have a dataset, and I want to randomly split it into two datasets. : For example, : Obs Policy # : 1 67 : 2 67 : 3 67 : 4 78 : 5 78 : .... : 10000000 77821178
|
g******7 发帖数: 1433 | 7 cluster? 能否具体说下code怎么写,多谢!
【在 D******n 的大作中提到】 : proc surveyselect;cluster;
|