Jay,

I VERY much appreciate you taking the time to so elaborately explain how
you go about merging.

Perhaps I should explain a little about my data. I basically have one
large database of over 55,000 records. A fairly substantial amount of it
I got from a cousin over 2 years ago. He has since add a fairly
substantial number (when he gave it to me he has 43,000 records. His
database is now over 63,000!)

So it would be reasonable to assume that somewhere close to his original
43,000 records I already have. Unfortunately when I do any variations of
merge I end up with SUBSTANTIAL messages to manually check. 

On a merge with "sounds like" use I end up with approximately 78,000
records to go through. With "exact" selected I still have to go through
39,000. 

I think you'll agree that I can't make the data I originally got only
better than the source it came from. Especially since in a lot of
instances the names are only related through marriage and I maintain
them only for the benefit of others who might be interested.

>From what I've noticed so far (and I've been staring at screens for a
fairly long time now), the best criteria for doesn't seem to be date if
anything else is different. In some instances there's a different spouse
and I can understand the logic that was used. However, in others what I
can see on the screen is sufficiently the same as to raise a question. 

For instance, the name of the individual is identical. So is the maiden
name of the mother but there's a second given name for the father.
Therefore I have to respond yes to the merge dialog box.

In a lot of instances, your suggestion for including estimated dates
won't satisfy an auto-merge. If you check the messages they indicate
that one of the reasons for a confirmation check is that the date is
approximate and not exact (ABT, BEF, AFT and EST would all qualify as
not exact wouldn't they?)

I do thank you for the time you took to explain how you merge. My query
was not how to merge but how the logic worked that the program used. I
was somehow hoping that somebody might be able to suggest something less
painful than responding Merge 39,000 times. *Sigh (I wonder what the
keyboard life is for Alt-M anyway? *Smiles)

Gary

-----Original Message-----
From: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED]] On Behalf Of J.M.
(Jay) Ingalls
Sent: Monday, November 19, 2001 12:02 PM
To: [EMAIL PROTECTED]
Subject: [LegacyUG] Merging genealogy records with Legacy

Gary,

I use the merge future a lot, but do not remember all the details of
how the AutoMerge decides what to merge on its own.

The basic premise is not to merge anything that is questionable!  That
should be your preference, also?

So, I select the "Merge", "Find Duplicates" options.

Then I do the backup.

For merging a large number of records:

On the merge options, I select "Special Searches".

Then I select an option that applies from that screen.

Using the "Using Ancestry Numbers Only " option, if the data came from
Ancestry files, this will do a rapid and fairly safe auto merge.
Records without enough data for Auto Merge to decide if they are the
same person will be left for me to do. Many matches will be obvious. I
do check parents, children, birth place, etc. I even use the "Edit"
option and check the sources sometimes. And, while at it, I edit the
data where appropriate.

Using the "Using ID Numbers Only " option, if the data includes ID
numbers that will be the same for each person, say data with Burleigh
numbers I have entered in files sent to other people, and they are
"coming back", multiple records with ID numbers like "B1" will be
presented, and merged if there are enough matches to meet the Auto
Merge criteria. If they are not merged by the program, then I must
review each one not merged, and decide if they are the same person.
Someone could have made an error.

As the merges are made, the program collects spouses and children of
the merged people, and presents them for review and merging as
appropriate.

I have not used the IntelliShare option.

After all these "Special" program proposed merges are merged or
rejected as not valid, I go to the "Normal Duplicate Search". I use
the reset button to return the settings to the default.

There I choose only the "Given Name" and "Check Birth Dates" options,
I turn off the "Include blank death dates" option. I leave the bottom
default menu settings.

I select the "continue" option, and precede.

After those merges are made, I add one more option at a time, starting
with the "Include blank death dates" option, and repeat the merging of
appropriate matches.

I have found one of the following options; "include blank burial
dates" or the "include blank death dates" locks up my computer. Do not
recall which one.

Automerge will pop up as it thinks necessary during these merges.

I do use the Deluxe? option to easily add any alternate names, dates,
and places to the events and notes as I merge records that I believe
are for the same person but have different data in some fields.

I use the "Not the same person" option to mark proposed merges that I
am sure are not the same person. This is a Legacy Deluxe option that
saves time on future merges, because they will not be proposed as
merges again.

During the merge process, I add estimated (abt) birth years to many
people, based on the birth year of the spouse, marriage date, and
birth dates of children. When I do that, I add a note to the persons
"General Note" saying what the "abt" date was based on, my name, and
the date. This can be a tremendous aide in identifying "matches" and
avoiding merging people born in the same place with the same names,
and sometimes with a wife named "Mary" or other common name, but they
are parent and child, or not related at all!

I do add "Jr." and "Sr." and "First" and "Second" as appropriate. Many
families had two children with the same name, because the first one
died. Some families used the same given (first) name for many
generations, so, in my Hill line, I have Israel I through V! All of
this added data helps provide data for my use and other researchers in
the future. Do note that two people with the same name living in the
same town were sometimes called "Junior" and "Senior" even though they
were not related! The same is true for "Big Joe" and "Little Joe" in
some towns or work crews.

Most importantly, I try to not create "bad data" that will mislead
other researchers.

Jay
============

>From: Gary Gratton <mailto:[EMAIL PROTECTED]>
>To: [EMAIL PROTECTED]
>Sent: Sunday, November 18, 2001 06:14 AM
>Subject: [LegacyUG] Merge Logic

>Here's a question for the group. What is the logic used by Millennia
on merging or more importantly auto-merging?
>Gary Gratton
===============
==============


To unsubscribe please visit:
http://www.legacyfamilytree.com/LegacyLists.asp

Legacy User Group Etiquette guidelines can be found at:
http://www.LegacyFamilyTree.com/Etiquette.asp

To unsubscribe please visit: http://www.legacyfamilytree.com/LegacyLists.asp

Legacy User Group Etiquette guidelines can be found at:
http://www.LegacyFamilyTree.com/Etiquette.asp

Reply via email to