NickL

My database appears to have developed a corruption in citations, which is potentially a disaster for me!  Looking for some guiding wisdom, please.

Using RM7.5.9.0 with a database of 5600 individuals, 377 sources, and until recently, about 25,000 citations.  Doing diagnostics for this problem, I notice that has now jumped to >29,000 citations.

I regularly run the Database Tools.  (Compact database does take noticeably longer now.)

I am well backed up, but the problem appears to have developed about 3 months ago, although I've only just noticed it.  Restoring to an anaffected version is almost unthinkable, given the amount of work that has been done since then.

Doing a LIST SOURCES report, I see many people/facts have developed additional "ghost" child citations.

For example, just picking one at random:

[SOURCE NAME] / 

..(first 50 citations not affected...)

..then...

51. EARWICKER, Liam-2199 (Birth).

52. EARWICKER, Liam-2199 (Birth). (Birth).

53. EARWICKER, Liam-2199 (Birth). (Birth). (Birth).

54. EARWICKER, Liam-2199 (Personal).

55. EARWICKER, Linda-29 (Birth).

56. EARWICKER, Rachel-2202 (Birth).

57. EARWICKER, Rachel-2202 (Birth). (Birth).

58. EARWICKER, Rachel-2202 (Birth). (Birth). (Birth).

..

Observations:

1. On-screen, everything appears correctly.  A Fact has a single instance of the citation.  The ghost-child citations only appear in a List of Sources.  (I haven't tried other reports though.)
2. These recursive 'ghosts' vary between three (3) as above, to commonly 12-15, to a couple of extremes of nearly 2000 ghost-children
3. The problem occurs on multiple (master) sources, but not all.
4. The problem is widespread, but not universal. Not all sources are affected. Not all citations from a particular source are affected.
5. Affected sources appear to come from different templates.
6. Changing the particular Fact citation (for example, adding or changing citation detail) and running the report again - the first / primary citation follows correctly, while the 'ghosts' remain unchanged. (So the ghosts were children at a point in time, and are not live mirrors of their parent.)
7. Deleting the particular citation from its Fact will make the parent and all ghost-children all disappear.  The ghost-children do not return when a new instance of the citation is added to the Fact.
8. The problem may be progressive, meaning more records may be affected as the database is worked.  I have not specifically investigated that though.
9. Doing a drag-and-drop of the whole affected dataset into a new empty database possibly resolves the problem. It appears to, but I haven't investigated that in detail, and I am afraid that I will lose something, but won't realise it until much later.

My dataset originated as an import from TMG about a year ago, and has been extensively re-worked within RM7 since then, with no problems.  I don't think this is an artifact of the import process.  I have done all work within the RM7 interface - there has been no external SQL manipulation.
 

Cheers

Nick