topic Re: Found Matching Values in single variable in SAS Programming

Found Matching Values in single variable

AnandsinhSindha — Tue, 08 Sep 2020 06:01:49 GMT

I need help to get my desire output through the sas coding.

I am using sas eg 8.1.

I have one data with column name "Name" and i want to found sound like name in same variable.

Data Have;

Name

mrityunjaya
gaurav
deepak
mrtunjay
dipak
piyush
mrityunjaya;

And i want output like this.(generate new column with "match" "non match" value)

Data want;

Name Flag

mrityunjaya match

gaurav not-match
deepak   match
mrtunjay   match
dipak match
piyush   not-match
mrityunjaya match

can some please help me how can i do this in sas.

Re: Found Matching Values in single variable

andreas_lds — Tue, 08 Sep 2020 06:47:45 GMT

Functions like spedis, compged and/or complev should help you to identify matches. The following step should get you started:

data have;
	length name $ 20;
	input name;
	datalines;
mrityunjaya
gaurav
deepak
mrtunjay
dipak
piyush
mrityunjaya
;

data scores;
	set have;

	do i = _n_+1 to numObs;
		set have(rename=(name=cmp)) nobs=numObs point=i;
		score_spedis = spedis(name, cmp);
		score_compged = compged(name, cmp);
		score_complev = complev(name, cmp);

		output;
	end;
run

Re: Found Matching Values in single variable

AnandsinhSindha — Tue, 08 Sep 2020 07:33:32 GMT

get some score based on above code but i am not getting my desired output....

Re: Found Matching Values in single variable

ChrisNZ — Tue, 08 Sep 2020 08:10:36 GMT

The answer to your question is:

proc sql;
  select unique a.NAME, max(a.NAME =* b.NAME) as FLAG 
  from TABLE a
         cross join
       TABLE b
  group by 1;

But a cross join gets really expensive really quickly when the volume of data increases.

10,000 records yield 100 million matches.

There are better ways, and this is typically an iterative process. For example you can remove the exact matches first to reduce the volume. Or you can try matching the first letter before using the sound like operator.

Re: Found Matching Values in single variable

andreas_lds — Tue, 08 Sep 2020 09:44:30 GMT

@AnandsinhSindha wrote:
get some score based on above code but i am not getting my desired output....

Of course not, i wrote: "The following step should get you started", meaning that it is up you to decide which function to use and which score is low enough to indicate a match.