topic Re: Match using names in SAS Programming

Match using names

abdulla — Sun, 17 Oct 2021 16:58:31 GMT

Hi,

I have attached an excel file. I want to match group A with group B using names. The different colors show the matching names. However, the names don't exactly match in two different groups. One of the ways we can match is by using the last name. But the last name is not always used similarly in the two groups, for example, John Holmes III. Also, in my large dataset, two different persons in the same firm and in the same year may have the same last name. How will I handle all these issues and how will I perfectly match one dataset with another dataset using names? Could you please help me?

Re: Match using names

Kurt_Bremser — Sun, 17 Oct 2021 19:51:12 GMT

Please post SAS datasets in usable form (working data steps with datalines).

Re: Match using names

SASKiwi — Sun, 17 Oct 2021 20:47:40 GMT

Perfect matching of free-text names is not possible. Only unique and totally accurate name keys like Customer_ID or Employee_ID will result in perfect matching.

The best you can do is to try and standardise the first and last names, if those are the only two columns available to you and try and get as high a match rate as possible. Also is it possible that two different people could share the same name?

Re: Match using names

PeterClemmensen — Mon, 18 Oct 2021 05:32:30 GMT

It will never be an exact match when the rules are not perfectly defined. However, for most purposes you can do a good job with fuzzy matching.

There are many axamples online .

Re: Match using names

ChrisNZ — Mon, 18 Oct 2021 07:34:31 GMT

> two different persons in the same firm and in the same year may have the same last name. How will I handle all these issues

Some questions such as these you need to decide what's teh best for you.

For fuzzy string matching, there are several SAS functions such as SPEDIS or COMPGED.