如何插入需要补全的值？

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

kind	value	AEBODSYS	AEDECOD
1	104
4	218
1	2	Blood and lymphatic system disorders
4	5	Blood and lymphatic system disorders
4	4	Blood and lymphatic system disorders	Anaemia
1	1	Blood and lymphatic system disorders	Hypochromic anaemia
1	1	Blood and lymphatic system disorders	Leukocytosis
4	1	Blood and lymphatic system disorders	Monocytosis
1	4	Cardiac disorders
4	6	Cardiac disorders
4	1	Cardiac disorders	Acute myocardial infarction
4	1	Cardiac disorders	Angina unstable
4	1	Cardiac disorders	Atrial fibrillation
4	1	Cardiac disorders	Atrioventricular block
4	1	Cardiac disorders	Cardiac failure
1	2	Cardiac disorders	Coronary artery disease
1	1	Cardiac disorders	Coronary artery insufficiency
4	2	Congenital, familial and genetic disorders
1	1	Congenital, familial and genetic disorders	Type IV hyperlipidaemia
4	1	Congenital, familial and genetic disorders	Type V hyperlipidaemia

我希望补全成

kind	value	AEBODSYS	AEDECOD
1	104
4	218
1	2	Blood and lymphatic system disorders
4	5	Blood and lymphatic system disorders
1	0	Blood and lymphatic system disorders	Anaemia
4	4	Blood and lymphatic system disorders	Anaemia
1	1	Blood and lymphatic system disorders	Hypochromic anaemia
4	0	Blood and lymphatic system disorders	Hypochromic anaemia
1	1	Blood and lymphatic system disorders	Leukocytosis
4	1	Blood and lymphatic system disorders	Monocytosis
1	4	Cardiac disorders
4	6	Cardiac disorders
1	0	Cardiac disorders	Acute myocardial infarction
4	1	Cardiac disorders	Acute myocardial infarction
1	0	Cardiac disorders	Angina unstable
4	1	Cardiac disorders	Angina unstable
1	0	Cardiac disorders	Atrial fibrillation
4	1	Cardiac disorders	Atrial fibrillation
1	0	Cardiac disorders	Atrioventricular block
4	1	Cardiac disorders	Atrioventricular block
1	0	Cardiac disorders	Cardiac failure
4	1	Cardiac disorders	Cardiac failure
1	2	Cardiac disorders	Coronary artery disease
4	0	Cardiac disorders	Coronary artery disease
1	1	Cardiac disorders	Coronary artery insufficiency
4	0	Cardiac disorders	Coronary artery insufficiency
1	0	Congenital, familial and genetic disorders
4	2	Congenital, familial and genetic disorders
1	0	Congenital, familial and genetic disorders	Type IV hyperlipidaemia
4	1	Congenital, familial and genetic disorders	Type IV hyperlipidaemia
1	0	Congenital, familial and genetic disorders	Type V hyperlipidaemia
4	1	Congenital, familial and genetic disorders	Type V hyperlipidaemia

就是每个AEDECOD或者AEBODSYS都有1,4两种kind；

有什么办法吗？

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0 回帖

关键词：Sufficiency Unstable Genetic disease Failure system 如何

提供一种可能的思路，不知道对不？
1.对原始数据集raw数据集sort by kind aebodsys aedecod nodupkey out=a;
2.建一个temp数据集（aebodsys aedecod为nodupkey的值）：do kind=1,4;do aebodsys=  ;aedecod= ;
3.merge raw temp;by kind aebodsys aedecod;
4.将value为空的值的row去掉

AEBODSYS='Blood and lymphatic system disorders', AEDECOD='Leukocytosis' 或 'Monocytosis'的时候，需不需要补？

就是补一条Leukocytosis的kind=4 和 Monocytosis的kind=1的？

wwang111 发表于 2016-8-26 16:25
AEBODSYS='Blood and lymphatic system disorders', AEDECOD='Leukocytosis' 或 'Monocytosis'的时候，需不 ...

要补的，这个是我疏忽了，简单来说，凡是出现过的记录，都得有1、4两条记录！

试试这个：

proc sql;
create table needcomp as
select distinct aebodsys, aedecod, kind
from test
group by 1,2
having count(*)=1;
quit;

data comp;
set needcomp;
if kind=4 then kind=1;
else if kind=1 then kind=4;
value=0;
run;

data newdata;
set test comp;
run;

proc sort data=newdata;
by aebodsys aedecod kind;
run;

wwang111 发表于 2016-8-26 16:40
试试这个：

proc sql;

我突然想到一种很另类的方法：
proc trance data=origin out=mid:
    by AEBODSYS   AEDECOD;
    var  value;
    id kind;
run;
data need;
    set mid(drop=_4 rename=(_1=vall));
    kind=1;
    set mid(drop=_1 rename=(_4=val));
    kind=4;
run;
是不是很巧妙？

这个方法没问题，只是你的代码错误太多，我给你改了一下。
transpose过程用到by的时候需要对by的变量先排序，如果原数据集的顺序不需要重新排列，那么可以加一个notsorted.

proc transpose data=origin out=mid(drop=_name_);
    by AEBODSYS AEDECOD notsorted;
    var  value;
    id kind;
run;

data need;
    set mid(in=a drop=_4 rename=(_1=value))
     mid(in=b drop=_1 rename=(_4=value));
         if a then kind=1;
         else if b then kind=4;
         if value =. then value=0;
run;

proc sort data=need;
by aebodsys aedecod kind;
run;

wwang111 发表于 2016-8-28 20:26
这个方法没问题，只是你的代码错误太多，我给你改了一下。
transpose过程用到by的时候需要对by的变量先排序 ...

好的。谢谢！

fyp198744 发表于 2016-8-28 07:03
我突然想到一种很另类的方法：
proc trance data=origin out=mid:
    by AEBODSYS   AEDECOD;

proc trance   是什么样的过程不，实现什么效果呀，还没见过呢

linshuhe123 发表于 2016-9-7 16:21
proc trance   是什么样的过程不，实现什么效果呀，还没见过呢

就是转置！