My RootsMagic database is sufficiently large that it seems impractical to open the whole thing in GenSmarts. GenSmarts can handle the whole thing, but I can't. There are just too many research suggestions, even with all the neat facilities that GenSmarts provides to winnow down the list of suggestions.
So I make a subset of my RootsMagic database, and open that in GenSmarts. My issue is that if I make a different subset of my RootsMagic database, and if the new subset contains some of the same people as the previous subset, I want GenSmarts to remember how how I already marked the research suggestions for those people.
GenSmarts seems extremely clever in how it does this, but I am curious about the details. ResearchResults.CSV seems to include the RootsMagic record number, the person's name, birth information, death information, a code for the research suggestion, and a code for how I have marked the suggestion (Found, Not Found, etc.). When GenSmarts starts up, it obviously has to match up the contents of my RootsMagic database (subset or otherwise) against the contents of ResearchResults.CSV. For a particular person, how much can the data differ and still have GenSmarts declare that the person in ResearchResults.CSV is the same person as the person in RootsMagic?
The answer to this question is important to my strategy for creating subsets of my RootsMagic database to feed into GenSmarts. For example, I can create subsets that preserve the RootsMagic record number, or that don't preserve the RootsMagic record number. How critical is that to the process?
On 2006-01-24 1:16pm GenSmarts replied...
A fuzzy match on name and personal details is what determines the match.
The RIN is simply used to expedite the matching process, and is only used if the file name is the same as it was when the result was recorded. If the RIN isn't used (or if there's a mis-match), it just takes a bit longer to sort out during the load. Wether you would notice the difference in the load time depends on the number of people loaded and the number of research results... We could, I suppose, let you overide this "use the RIN" logic by putting in a setting that said to use the RIN if the file names were "sufficiently" similar.
A better solution, of course, is just to tell GenSmarts what portion of your primary file you want to analyze. We have a beta of this "partial analysis" functionality that will be in an upcoming release. Email us at email@example.com if you want to take a look at the beta.