topic Re: Reg :Duplicates in SAS Procedures

Reg :Duplicates

R_Win — Tue, 31 Mar 2009 06:17:46 GMT

i have a data set

Data m;
input id name $;
cards;
1 Raju
2 nani
2 nani
2 kool
3 india
4 usa
4usa
4usa
5 uk
run;

Now i want the out put into other dataset with out coming of one duplicates to another dataset.the output shd be like this

id name
1 Raju
2 kool
3 india
5 uk

As the 2nani, and 4,usa are alredy repeted that observations shd not come to the new dataset
Message was edited by: Main Message was edited by: Main

Re: Reg :Duplicates

deleted_user — Tue, 31 Mar 2009 07:50:56 GMT

Hi,

u will get the desired output by using the following code.

Data m;
input id name $;
cards;
1 Raju
2 nani
2 nani
2 nani
3 india
4 usa
4 usa
4 usa
5 uk
run;
proc sort data=m;
by id;
run ;

data m_new;
set m;
by id;
if first.id ne 1 or last.id ne 1 then delete;
else if first.id then output;
run;

Re: Reg :Duplicates

sbb — Tue, 31 Mar 2009 07:54:33 GMT

After sorting your data file in desired order with a BY variable list, then use a DATA step with the two statements IF FIRST. and also IF LAST. to identify totally unique one-occurence data conditions, and output those to one file, while also outputting the others to a separate file.

Have a look at the SAS support http://support.sas.com/ website and its SEARCH facility, to search on the words:

duplicate by variable

and you will find both SAS product documentation and technical reference papers on the topic.

Scott Barry
SBBWorks, Inc.

Re: Reg :Duplicates

GertNissen — Tue, 31 Mar 2009 07:56:40 GMT

proc summary data=m n;
by id name;
output out=freq(where=(_FREQ_=1);
run;

Re: Reg :Duplicates

ssas — Wed, 01 Apr 2009 13:25:59 GMT

HI,
why can't you try this

proc sort data=a nodupkey out=b;
by id name;
run;

data c;
set b;
if first.id eq 1 then output c;
run;

or

a simple way

proc sql;
create table c as select distinct id,name from b;
quit;

hope it may help you
sams

Re: Reg :Duplicates

deleted_user — Wed, 01 Apr 2009 20:13:27 GMT

as your data appear to be approximately in order, you could try[pre] data reduced ;
set m ;
by id NOTSORTED ;
if last.id ;
run ;[/pre]This returns the last row in the current order, for each ID.

PeterC