Jump to content


Photo

File compare not matching


  • Please log in to reply
11 replies to this topic

#1 JeffH13

JeffH13

    Advanced Member

  • Members
  • PipPipPip
  • 154 posts

Posted 21 March 2015 - 06:22 AM

I'm not having much success with the file compare feature.

 

Why do I get 0% match on these entries in the two files I'm comparing?  Because of that, the two don't line up next to each other in file compare so I can't copy facts from one file to another. 

 

                  File 1                                                   File 2

Person      Herman Hedberg                    Herman August Hedberg

Name        Herman Hedberg                    Herman August Hedberg

                                                                 August Herman Hedberg (Alt Name fact)

Birth          27 Mar 1882 Sweden              27 Mar 1882 Sweden

Father       Andrew Hedberg                     Andrew Hedberg

Mother      Emma Stromberg                    Emma Stromberg

 

File 2 has some additional facts for this person that File 1 doesn't such as Naturalization, Immigration, Occupation, Residence.  They do not have the same RIN.

 

I would expect some % greater than zero.

 

Thanks,

Jeff



#2 Vyger

Vyger

    Advanced Member

  • Members
  • PipPipPip
  • 3434 posts

Posted 21 March 2015 - 07:37 AM

I'm not having much success with the file compare feature.

 

Firstly I am not using RM7 even though this is the one feature addition which interested me. Therefore I cannot speak first hand but for my own interest I have been reading all posts pertaining to File Compare with interest.

 

It has been described in many ways (DYOR) but basically it looks not to be the tool users were expecting or wishing for, I'm sure you will get other comments. 


Customers should never be frustrated by things they cannot do.

 

User of Family Historian 6.2.7, Rootsmagic 7.6.2, Family Tree Maker 2014 & Legacy 7.5

 

Excel to Gedcom conversion - simple getting started tutorials here

 

Root


#3 Laura

Laura

    Advanced Member

  • Members
  • PipPipPip
  • 4276 posts

Posted 21 March 2015 - 04:47 PM

I set up two trial databases for this person and his parents.  I typed the data given above into both databases and didn't drag and drop from one database to the other so the UID's would not be the same.

Database Trial1 has the birth date entered.  Database Trial 2 has the birth date and Alternate name entered.  I didn't add any other facts to Trial2.

I opened Trial1 and compared Trial2.

This is the results in the top pane:

0%     Trial1 - [Blank]                           Trial2 - Hedberg Andrew -1

0%     Trial1 - [Blank]                           Trial2 - Stromberg, Emma-2

0%     Trial1 - Hedberg, Andrew-2.    Trial2 - [Blank]

76%   Trial1 - Hedberg, Herman-1    Trial2 - Hedberg, Herman August-1

0%     Trial1 - Stromberg, Emma-1.   Trial2 - [Blank]

Did you check the matches or just the 0%.

#4 Laura

Laura

    Advanced Member

  • Members
  • PipPipPip
  • 4276 posts

Posted 21 March 2015 - 09:17 PM

I was wondering about what happened to color coding when copying a record from the compared database to the main database.

So, I copied my main database, Database1, and copied the database to compare, Database2.

I cleared the color coding for everyone in Database1, Black, and color coded everyone in Database2 as Red.

The people that I copied from Database2 to Database1 retained the Red color coding.

I was copying people from Database2 with 0% matches with Database1. Many of them are people who are in both databases but the names are different so they aren't matched and the UID's are different.

This is very helpful as I can then open Database1 and Database2 side by side and check if I already have that person in Database1 with a different name.

Since I doubt if I will ever transfer these people directly to my actual main database, I can change the color coding of the people I do want to transfer to another color in Database1 and clean up their places and sources, etc. before transferring them into the actual main database. After the transfer, I can change the people I transferred to yet another color in Database1 to help keep track of what I have done.

#5 JeffH13

JeffH13

    Advanced Member

  • Members
  • PipPipPip
  • 154 posts

Posted 21 March 2015 - 09:42 PM

First of all, in my first post, I've correct "UID" to "RIN".  So below, following each name I have the RIN just as RM shows it.

Laura, I'm not sure what you mean "Did you check the matches or just the 0%"

Anyways, I created 3 new data bases, each with the same three people: Andrew Hedberg, Emma Stromberg, and son Herman Hedberg. The parents have a marriage fact but no date or place. There are no other facts: no birthdates, no alternate names, nothing.

Database1 entered child, then father then mother

Database2 entered father, then mother, then child

Database3 entered father, then mother, then child (exact same as Database 2)

First comparison (same data, different order of entry):

% Match          Database2                 Database1

0                                                        Hedberg, Andrew-2
0                                                        Stromberg, Emma-3
0                     Hedberg, Andrew-1
98                   Hedberg, Herman-3   Hedberg, Herman-1
0                     Stromberg, Emma-2


Second comparison Database2 and Database3 (same data, same order of entry): 100% match on all 3 persons.


Returning to the first comparison and adding Andrew's exact birthdate 9 Apr 1854 to both Database 1 & 2 I get:

% Match        Database2                   Database1
0                                                        Stromberg, Emma-3
79                  Hedberg, Andrew-1     Hedberg, Andrew-2
98                  Hedberg, Herman-3    Hedberg, Herman -1
0                    Stromberg, Emma-2

Why is the Andrew-Andrew match only 79% especially since they both have the exact same birth date, and Herman-Herman at 98% has no birth entered?

Repeating this except Andrew's birthdate in Database2 is 9 Apr 1854 and in Database 1 is just 1854:

I get the same results as the "First Comparison" above. That is 98% for Herman-Herman (no birthdates) and 0% for Andrew (similar birthdates) and Emma.

I would have thought that comparing 1854 vs 9 Apr 1854 for two persons (Andrew in both databases) with the same name would produce >0% match.

So there seems to be some disfunctionality.

 

Jeff
 



#6 Renee Zamora

Renee Zamora

    Advanced Member

  • Support
  • PipPipPip
  • 8491 posts

Posted 23 March 2015 - 09:18 AM

If the person has multiple facts of the same type or alternate names each needs to be compared against each other. Even if they have exactly the same data, the comparison between the same type facts will lower the match percentage. 

 

If a person has a similar name and same birth date they won't necessarily be considered a match. Having twins born into a family is a good example. I think that is why compare file was having an issue with Herman Hedberg. The 2nd database used August in the name, even starting with that as an alternate name. If the first database had the August even as an alternate name it would have probably considered them a possible match. 


Renee
RootsMagic

#7 TomH

TomH

    Advanced Member

  • Members
  • PipPipPip
  • 6268 posts

Posted 23 March 2015 - 10:47 AM

A useful enhancement would give the user control over how tightly each comparison should be in order for a possible match to be offered by the program.

Tom user of RM7550 FTM2017 Ancestry.ca FamilySearch.org FindMyPast.com
SQLite_Tools_For_Roots_Magic_in_PR_Celti wiki, exploiting the database in special ways >>> RMtrix-tiny.png app, a bundle of RootsMagic utilities.


#8 Renee Zamora

Renee Zamora

    Advanced Member

  • Support
  • PipPipPip
  • 8491 posts

Posted 23 March 2015 - 12:05 PM

Settings like Duplicate Search Merge then?


Renee
RootsMagic

#9 TomH

TomH

    Advanced Member

  • Members
  • PipPipPip
  • 6268 posts

Posted 23 March 2015 - 12:56 PM

That is one approach and it would be consistent. Whether there is a better one that could be applied to both File Compare and Duplicate Search is moot - there is the Ancestry.com sliders... 


Tom user of RM7550 FTM2017 Ancestry.ca FamilySearch.org FindMyPast.com
SQLite_Tools_For_Roots_Magic_in_PR_Celti wiki, exploiting the database in special ways >>> RMtrix-tiny.png app, a bundle of RootsMagic utilities.


#10 Vyger

Vyger

    Advanced Member

  • Members
  • PipPipPip
  • 3434 posts

Posted 23 March 2015 - 01:27 PM

A useful enhancement would give the user control over how tightly each comparison should be in order for a possible match to be offered by the program.

 

Settings like Duplicate Search Merge then?

 

I have been thinking about this and not being a user of RM7 was going to watch the video again but I now assume such settings do not exist.

 

I am well aware of the quirks of DSM and have my opinions, most of which I have shared, regarding how I believe it sould be improved. Name soundex does have to be used where middle names exist on one side only, various other options have to be employed to widen the match to include what to the eye look very obvious matches.

 

Unfortunately widening the scope of DSM on large databases provides a list which becomes almost prohibitive whereas on a small database it can be very useful. The Start with Surname option was a little help some versions ago but I believe utilizing Named Groups and restricting the scope of DSM to within that Group (subset) would be very advantageous. So Instead of Start with Surname you could have Use Named Group.

 

One could build a Group of Surname > Sounds Like & Any Place > Contains or whatever else suits and then widen the scope of DSM to include Sound alike names and widening the Birth and Death year span but still producing a smaller list of possible matches.

 

I find DSM an extremely useful tool and I am producing a short video to illustrate it's usefulness and what I judge to be some of it's shortcomings.


Customers should never be frustrated by things they cannot do.

 

User of Family Historian 6.2.7, Rootsmagic 7.6.2, Family Tree Maker 2014 & Legacy 7.5

 

Excel to Gedcom conversion - simple getting started tutorials here

 

Root


#11 JeffH13

JeffH13

    Advanced Member

  • Members
  • PipPipPip
  • 154 posts

Posted 23 March 2015 - 05:38 PM

Renee, please look at my second post 21 March, 9:42PM.

There has to be >0% match in the examples in my two posts. I have several files started at different times that I'm trying to compare and I'm mostly getting 0%. (The files contain many of the same persons.) If the matches were >0% then I could copy facts etc. between files.

If nothing else, what if the user could manually match, then the compare function would allow copying facts, sources, etc.

Thanks,
Jeff

#12 Renee Zamora

Renee Zamora

    Advanced Member

  • Support
  • PipPipPip
  • 8491 posts

Posted 24 March 2015 - 01:05 PM

On your example I think its simply because the names are different. That's why its not a match. If you want to make it a match then add the August to the other person's name and it should show up as a match. You have to make the person more of a match to have it show as a potential match. 


Renee
RootsMagic