I have two files, one is really screwed up, and the other is in good shape. When I say screwed up, it lists voters with the wrong precinct, legislative district, senate district, and congressional district. Its literally like somebody threw all the voters into a hat and jumbled up thier information.

What I would like to do is compare the correct voter file with the bad voter file and determine what percentage of the database is correct. The result will be something as simple as:

Match = 85%

The definite match in each file is address & city. What is questionable is the Precinct, Leg District, Senate District, and Congressional District.

Also, both files are in FM8 format. I assume the first step is create a relationship between the files, but I'm fuzzy on which fields and ultimately getting percentage of successful match.

An example dataset is below

Address: 123 East Smith Street

City: Yourville

Precinct: 2397

LegDistrict: 22

SenDistrict: 7

CongDistrict: 3

Thanks in advance...

Them voters again? Yes, you will probably need a relationship. But first, two questions:

1. Does the combination Address & City together identify a record UNIQUELY in both files?

2. Are these two files the same otherwise? IOW, if we remove the district information from both, will they be essentialy copies of one another?

Assuming the answer is yes to both, you can define your relationship as:

CorrectFile::Address = BadFile::Address

AND

CorrectFile::City = BadFile::City

Next, define a calculation field cIsCorrect in the BadFile (result is number) =

Precinct = CorrectFile::Precinct and

LegDistrict = CorrectFile::LegDistrict and

SenDistrict = CorrectFile::SenDistrict and

CongDistrict = CorrectFile::CongDistrict

Also define a summary field sCorrect as Total of cIsCorrect.

Finally, a calculation field =

sCorrect / Get ( FoundCount )

will return the fraction of correct records in the found set of the BadFile.

Comment,

Once again you pulled my rear end from the fire.

Thanks for help.