How to match COMPGED scores to 100 or less?

[Smile1]

Hi~

To match COMPGED scores to 100 or less, I used compged function as below.

==========================================

libname proj "D:\SASWORK";
proc sql; *evaluate name similarity with score 0~100;
create table nameexample
as select *, compged (lender, mgrname) as similarity
from proj.periodexample
where calculated similarity le 100;
quit;

==========================================

There is no problem from line 1 to 5.

However, when I add 'where calculated similarity le 100;', I can't get the result.

The log message is as follows : 

=========================================================

Note : Table WORK.NAMEEXAMPLE created, with 0 rows and 11 columns.

=========================================================

I would really appreciate it if anyone could give me a help.

Thanks!

[sbxkoenk · Accepted solution]

Hello @Smile1 ,

I guess you do not have any COMPGED costs below (ore equal to) 100.

Please check for minimal 'similarity' value after running your PROC SQL without the where clause.

Or maybe you need some COMPGED modifiers (to ignore case, to remove blanks, to remove quotation marks during / before comparison).

If there are WARNINGs or ERRORs in your LOG, please publish your LOG in this thread.
Make use of the Insert Code icon </> in the toolbar and paste your LOG in the pop-up window. That way the LOG preserves structure and formatting (and some colors).

Thanks,

Koen

[Smile1]

Hi Koen,

Thank you so much for your help! I added modifier 'I' as you advised.
I coded as below and my sas program was finally working. Thank you so much! 🙂

==========================================
libname proj "D:\SASWORK";
proc sql; *evaluate name similarity with score 0~100;
create table nameexample
as select *, compged (lender, mgrname, 'I') as similarity
from proj.periodexample
where calculated similarity le 100;
quit;
==========================================

[PGStats]

Note: The main advantage of COMPGED is the ability to modify the costs of edit operations by calling CALL COMPCOST prior to calling the COMPGED function. But you can't do that in SQL... so you might be better using the COMPLEV function instead. It is faster and the results are easier to understand.

complev (lender, mgrname) le 2;

[Smile1]

Thank you for your advice, PG. Now, my COMPGED function is working. I will try to use COMPLEV function as well sooner or later.

Thanks, have a wonderful day! 🙂