I have encountered on several occasions that when creating gedcoms from a database after deleting individuals in a group, or by selecting a group for export which excludes records in a 2nd group, the gedcom is often corrupt. Reading the corrupt Gedcom into a text editor shows what appears to be a line of null characters breaking the body of the gedcom. The Gedcom cannot be read by FTAnalyser returning the message;
Invalid GEDCOM, Line: 103449: 'Invalid GEDCOM, Line: 103450: ‘5 NAME Folio’. Last Name Seen: Gwenda Wilmot /Fredericks/. Error was: Level numbers must increase by 1
This issue may affect any database subsets used to create Gedcoms.
Thanks. It’s obvious the .GED file is corrupted. Those vertical rectangle characters prevent usability of the file for subsequent Import anywhere. Are there any other instances of those characters?
You should definitely run Tools>Database tools (Integrity check, Rebuild indexes, Remove phantom records, Compact database)
Now, export an RM GEDCOM that includes just Gwenda and her family (both birth and marital). Ensure the .GED no longer contains those characters. Run the analyzer.
If all is good …repeat the process with the full complement of folks as the export. Good Luck!
There also seems to be issues with links between records when a part of a family is not included in the group. Sample from FTAnalyser
Child not found in family: F21: William Henry Allsopp and Sarah Davis
Child not found in family: F22: Samuel Henry Allsopp and Ellen Elizabeth Sullivan
Child not found in family: F22: Samuel Henry Allsopp and Ellen Elizabeth Sullivan
Child not found in family: F187: Elijah John Allsopp and Mary Ann Cornish
Child not found in family: F235: Jack Malden Allsopp and Marie Valerie Hummelstad
Child not found in family: F286: Frederick William Allsopp and Hannah Levington Claughton
Child not found in family: F427: Frederick George Allsopp and Ethel May Barnes
I have run the Database > tools multiple times. First port of call. I can generate a gedcom without issue for the entire database, a tree, or family. It is with using groups where subsets are selected.
Those .GED listings are obviously corrupted, looking almost like non-UTF8 text.
Do you mean Marking some Group(s), then Unmarking some people (or Groups) to get a subset? I’m not seeing the problem, but am only doing the “obvious” methods.
Are you on an Apple device?
What has changed since it was last working? Prior version of RM? Prior version of computer operating system?
@dennis – not able to help you with the problem BUT I am wondering, can you drop and drag your group to a new database and then make a gedcom–perhaps that would tell you something…
Sorry to be a nuisance but I think I found the problem. The Gedcom files were being output to a synched folder on my NAS and it is possible that the synchronisation was occurring during the creation of the file causing an interruption to the output. When output to a non-synched folder all went smoothly.
Thanks to all for your feedback… we learn something new everyday.
I would first rule out the GEDCOM before looking for issues with database (at least after you run all tools and making sure the database is not being sync in cloud).
