怎样用sql把一个数据集纵向截成三段然后横向连接，自己编了一个，有error

是 否 +2 论坛币

k人 参与回答

经管之家送您一份

应届毕业生专属福利!

求职就业群

赵安豆老师微信：zhaoandou666

经管之家联合CDA

送您一个全额奖学金名额~ !

立即领取

感谢您参与论坛问题回答

经管之家送您两个论坛币！

+2 论坛币

这个数据集

要按照group截成三段，然后横向连接，成这样，用sql'做

这个数据集是我用别的方法做的

1. data means_2 (drop=Group);
   
2. merge means_1 (where=(group="A") rename=(age=GroupA))
   
3. means_1 (where=(group="B") rename=(age=GroupB))
   
4. means_1 (where=(group="C") rename=(age=Total)) ;
   
5. run ;

复制代码

自己编了一个sql的版本，第一个数据集是means—1

1. proc sql ;
   
2. create table work.means_2 as
   
3. select means_1.*
   
4. from work.means_1 (where=(group="A"))
   
5. full join work.means_1 (where=(group="B"))
   
6. on means_1._STAT_=means_1._STAT_ ;
   
7. quit ;

复制代码

有error

请教各位大神
这是哪里出现问题了
该怎么改正，
感激不尽

扫码加我 拉你入群

请注明：姓名-公司-职位

以便审核进群资格，未注明则拒绝

分享0 收藏1 回帖

关键词：Error 数据集 sql err proc sql error

you should give two different aliases to the same table

from work.means_1 (where=(group="A")) as A
full join work.means_1 (where=(group="B")) as B
on A._STAT_=B._STAT_ ;

为什么一定要用SQL呢，既然有更简单的方法。

do循环按ABC截取3个数据集，然后merge不好么

teqel 发表于 2015-9-22 01:05
from work.means_1 (where=(group="A")) as A
full join work.means_1 (where=(group="B")) as B
on A._S ...

81   proc sql ;
81 !                                                       *xxxxxxxxxxxxxxxxxxxxxxxx;
82        create table work.means_2  as
83           select  means_1.*
84           from work.means_1 (where=(group="A")) as A
85           full join work.means_1 (where=(group="B")) as B
86           on A._STAT_=B._STAT_  ;
WARNING: Correlation name means_1 identifies more than one table.
WARNING: Variable Group already exists on file WORK.MEANS_2.
WARNING: Variable _STAT_ already exists on file WORK.MEANS_2.
WARNING: Variable Age already exists on file WORK.MEANS_2.
NOTE: Table WORK.MEANS_2 created, with 5 rows and 3 columns.

有错误啊  数据集只出来了B列

godxh 发表于 2015-9-22 14:29
为什么一定要用SQL呢，既然有更简单的方法。

81   proc sql ;
81 !                                                       *xxxxxxxxxxxxxxxxxxxxxxxx;
82        create table work.means_2  as
83           select  means_1.*
84           from work.means_1 (where=(group="A")) as A
85           full join work.means_1 (where=(group="B")) as B
86           on A._STAT_=B._STAT_  ;
WARNING: Correlation name means_1 identifies more than one table.
WARNING: Variable Group already exists on file WORK.MEANS_2.
WARNING: Variable _STAT_ already exists on file WORK.MEANS_2.
WARNING: Variable Age already exists on file WORK.MEANS_2.
NOTE: Table WORK.MEANS_2 created, with 5 rows and 3 columns.

数据集只出来了GroupB的，

teqel 发表于 2015-9-22 01:05
from work.means_1 (where=(group="A")) as A
full join work.means_1 (where=(group="B")) as B
on A._S ...

81   proc sql ;
81 !                                                       *xxxxxxxxxxxxxxxxxxxxxxxx;
82        create table work.means_2  as
83           select  means_1.*
84           from work.means_1 (where=(group="A")) as A
85           full join work.means_1 (where=(group="B")) as B
86           on A._STAT_=B._STAT_  ;
WARNING: Correlation name means_1 identifies more than one table.
WARNING: Variable Group already exists on file WORK.MEANS_2.
WARNING: Variable _STAT_ already exists on file WORK.MEANS_2.
WARNING: Variable Age already exists on file WORK.MEANS_2.
NOTE: Table WORK.MEANS_2 created, with 5 rows and 3 columns.

数据集只出来了GroupB的，

godxh 发表于 2015-9-22 14:29
为什么一定要用SQL呢，既然有更简单的方法。

想尝试一下别的方法