topic Re: Separating first name and last name in SAS Programming

Separating first name and last name

gurpreetkaur — Thu, 05 Nov 2015 03:37:45 GMT

Hi Experts,

I need to separate first name and last name which do not have any delimiter specified.

I have the following data set:

data sample;
input name : $20.;
datalines;
gurpreetkaur
anujmadan
dhirajdhingra
mehakbhasin
sahibakhurana
;
run;

I need a dataset from the above dataset with ouput as:

first name	last name
gurpreet	kaur
anuj	madan
dhiraj	dhingra
mehak	bhasin
sahiba	khurana

how should I get the above data set??

Thanks !!

Gurpreet Kaur

Re: Separating first name and last name

SASKiwi — Thu, 05 Nov 2015 04:15:31 GMT

If there is no rule or pattern you can describe that identifies where the word boundaries are then there is no way to program it in SAS or any other computer language.

The only way I suspect this would work is if your software had access to an appropriate database of surnames and first names and you were able to do name matches. This is definitely not available in Base SAS.

Re: Separating first name and last name

Jagadishkatam — Thu, 05 Nov 2015 11:18:38 GMT

I agree with saskiwi on this. We need a pattern or some information which helps differentiate the surname and last name

Re: Separating first name and last name

DartRodrigo — Thu, 05 Nov 2015 13:51:55 GMT

Hi mate,

For this small table you can use multiple IFs and substr function to separate what you need.

But if your table is very large and there is no patterns, then you need a new way to get this table

I agree with Kiwi.

But for this small table you can do this:

%macro name(name,pos,len,pos1,len1);
 
 data want_&name(where=(last_name ne ""));
 set sample;
 if substr(name,1,1) = "&name." then do;
 last_name = substr(name,&pos. ,&len.);
 first_name = substr(name,&pos1.,&len1.);
 end;
 run;

%mend;

%name(g,9,6 ,1,8);
%name(a,5,6 ,1,4);
%name(d,7,10,1,9);
%name(m,6,10,1,5);
%name(s,7,10,1,6);

And then append.

Hope this helps

Re: Separating first name and last name

MaikH_Schutze — Thu, 05 Nov 2015 14:41:08 GMT

Agree with everyone above. Even something as simple as capital letter for each first and last name would help. Anything to determine the start position of each name element (including middle initials if there are any).

So if you altered the data in any way to make them all lowercase, then maybe you have the information you need, otherwise, you will need to request a new file with the appropriate delimiters.

Re: Separating first name and last name

gurpreetkaur — Fri, 06 Nov 2015 06:18:37 GMT

Thank you all for your replies. I was asked this question in one of my interviews. I answered the same but the interviewer was not convinced and then I thought there may be a way to do this.

Re: Separating first name and last name

ballardw — Fri, 06 Nov 2015 15:36:00 GMT

It is possible that what you presented was not exactly what the interviewer asked or the data example you provided had some difference.

A very minor change such as the name parts starting with a capital letter would make it possible.