SAS的联盟问题

Question

我遇到了关于SAS联盟的问题。

1）当我运行以下代码时：

proc sql;
select count(*) as cnt_obs, count(distinct user) as cnt_user from input1
union all
select count(*) as cnt_obs, count(distinct user) as cnt_user from input2
union all
select count(*) as cnt_obs, count(distinct user) as cnt_user from input3
union all
select count(*) as cnt_obs, count(distinct user) as cnt_user from input4
quit;

我得到了结果：

cnt_obs    cnt_user
10000000   4983437
9771110    4983438
8345177    4983438
8188069    4983438

2）但是根据我在proc sql步骤之前的过程，结果的第二列应该是相同的，所以我运行另一个代码来检查我的结果。当我运行以下代码时：

proc sql;
select count(*) as cnt_obs, count(distinct user) as cnt_user from input1;
select count(*) as cnt_obs, count(distinct user) as cnt_user from input2;
select count(*) as cnt_obs, count(distinct user) as cnt_user from input3;
select count(*) as cnt_obs, count(distinct user) as cnt_user from input4;
quit;

我得到了预期的结果：

cnt_obs    cnt_user
10000000   4983437
9771110    4983437
8345177    4983437
8188069    4983437

3）在SAS proc sql步骤中UNION似乎存在问题，有谁知道问题是什么？非常感谢！