Set Relationship - problems

note-- I actually moved my info from another post as I have a few other issues and comments BUT did NOT want to take over someone else’s post–

Tom-- it has nothing to do with your database or what you’ve done to it as I am seeing the same issue in several of my databases-- some strictly made in RM and others started in other genealogy programs and moved thru several different programs until finally getting to RM 7 and all the various versions…
So my dad’s database and one other seems to have no problems ( but the 2nd one is so large, it might take years before I find a problem)…
Mom’s database seems to have 1–and it’s bad as it’s just one of my sis-in-laws–brother and her kids show the correct relationships— so I tried dragging and dropping my sis-in-law to dad’s database and it shows the correct relationship-- Interesting!!!
Then 1 of hubby’s databases has a few missing the relationships of spouse of the g-uncles (2 out of 4) and spouse of gg-aunt/gg-uncle ( 1 out of 7) that I have found so far–again all the g-uncles and gg-aunts/ gg-uncle show correctly-- their are some g-uncles who were married 3 or more times–2 of the wives show as spouse of-- the 3rd doesn’t…
Then we come to my test database and it is a mess–it’s like every generation is missing at least one ( and sometimes more) spouse of relationships ----a lot of these were originally entered into the RM 9 database BUT I have also added new individuals into the RM 10 database–set the relationship again to the same person to update the file and some show the spouse of relationship and some don’t…
Going back to I dragged and dropped my sis-in-law and it corrected it–well I dropped and dragged my whole test database–well it corrected some BUT not all – when I added some new people–some showed correctly–some don’t BUT the worse part is that some who show right in the original RM 10
test database
image

don’t show right in the drag and drop test database
image
one step forward-- 2 steps backwards!!!

Tried everything I could think of to fix the problem up to unlinking brother, sis-in-law and their kids from each other and unlinking brother and sis-in-law from their parents and then relinking them all and resetting set relationship-- no change…

Edit–on sis-in-law–I noticed before I unlinked everybody and it’s the same after relinking everybody-- when I set the relationship to myself–sis-in-law Shirley does NOT show as my sister-in-law ( so therefore NOT in my Kinship Report either)–when I set relationship to Shirley, it shows me as her sister-in-law–so I would show up on her Kinship List…

That definitely should NOT happen as the only way we connect is thru marriage…

RM 10.0.0.0 windows 11-- not synching to any cloud service–no dual monitors

2nd issue I noticed on one of my databases where the “spouse of” relationships were right ( by right I mean showing)…
To explain the issue, my hubby’s 3rd g-grandma Louise Green married Geo Ketchum and Louise’s brother Squire Green married Eliza Ketchum ( siblings marrying siblings) BUT hubby only descends from Geo and Louisa–
so hubby is a descendant of Louisa and Squire’s parents James and Margaret Green
and he is a descendant of Geo and Eliza parents James and Mary Ketchum
( trying to show there are no 1st cousin marriages in this group that lead back to the 1 same set of ancestors)

So when I use set relationship to my hubby-- Eliza’s relationship is correct
image
But NOT Squire’s relationship–it should say 3rd great-granduncle
image

If I go in and set relationship to Squire, it says hubby is his 3rd great-grandnephew–which is correct…

In another thread @

So this should NOT happen–as there is one line of descent thru the Green and a second thru the Ketchums…

Thanks for splitting away from the enhancement request this problem that RM10 Set Relationships gives unreliable and inconsistent results for spouses of blood relatives. This affects not only the extension of ancestors’ spouses beyond RM9’s range but also those that were in RM9’s range, as close as uncle/aunt.

I, too, did a couple of transfers to see if that resolved my missing aunt relationship with the following findings:

  1. A plain GEDCOM export-import resolved that issue
  2. A drag’n’drop did not. Note that RM10 no longer uses a background GEDCOM export-import for drag;n;drop.

Both were selective transfers, just the descendants and spouses of my grandfather, each to a new empty database. #1 was as plain as you can get - no options selected. I haven’t repeated with RM-specific details included but I doubt that will have an effect - export uses a procedure quite different from Set Relationships to mark people.

I don’t yet completely understand the problem, but I have now spent enough time with it that I can demonstrate some symptoms. Here are a series of screen shots that show a problem - perhaps not “the” problem, but certainly “a” problem. My color codes have not been regenerated since RM10, but my Set Relationships have been regenerated since RM10. I will not be able to regenerate my color codes in RM10 until I have a better understanding of the problem. Any color codes you see were generated in RM9 and have simply been retained in RM10. Fortunately, I don’t have any immediate need to regenerate my color codes in RM10.
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
First we highlight Nancy Elizabeth (Lizzie) Bryan without a problem. She was indeed my second great-grandaunt. She is color coded red, indicating she is a relative.


+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Next we highlight her husband Thomas Simeon (Tom) Roberts. In RM9, he showed blank for the relationship because RM9 did not support spouse of relationships. In RM10, he still shows blank but he actually should show spouse of second great grand-aunt because RM10 does support spouse of relationships. This seems to be a bug. And indeed for the vast majority of my spouse of people, RM10 indeed does show show them as spouse of. So I wonder why Tom Roberts is not a spouse of person. My color coding from RM9 shows him as green, which is my color code for spouse of. My color coding criterion for green is “spouse of a red person and not yet colored red”.

+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
Next we move Tom Roberts into the primary position and see clearly that he has two spouses. The aforementioned Nancy Elizabeth (Lizzie) Bryan is still red. The newly visible Margaret R. Lawson is color coded blue. My criterion for color coding blue people is “spouse of a green person and not yet colored red or green”. I question whether I even want to keep the blue people in my database, but so far I do keep all the blue people in my database.

+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
My next guess was that the problem was that Tom Roberts had two spouses. So I made a test database. In the test database I deleted his second spouse and reran Set Relationships. I actually reran Set Relationships twice, including clearing the relationships before setting them again. Tom Roberts is still not a spouse of. The vast majority of my spouse of people actually show as spouse of. But not Tom Roberts. I have no explanation and I see no pattern for why Tom Roberts was not marked spouse of. But clearly there is a bug.

+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
My color coding in RM9 was accomplished using an SQLite script. The script looked at two pieces of data. First, it looked at the relationship indicators as set by RM9’s version of Set Relationship. It didn’t care what the relationship was, just that the relationship was blank or not blank. So a not blank relationship caused the script to color code the person as red. Second, it looked at RM9’s FamilyTable and ChildTable to determine relationships. Therefore, the script did not have to interpret any of RM9’.s internal codes for relationships beyond blank and not blank. In order to adapt the script to RM10, the script will need to interpret RM10’s internal codes for relationships. But I can’t do that until the script sets the relationship correctly for people such as Tom Roberts. In the meantime, my color codes from RM9 are still fine in RM10.

Another observation about RM10 Set Relationships is that it is not symmetrically reporting relationships. For example, in my cleaned by export-import, basic 78 person database of descendants from one of my grandparents, my two aunts are so labelled when I am the reference but I am not labelled as their nephew when either of them is the reference. Now that may be perfectly reasonable - aunts by marriage are likely to have a greater influence on their spouse’s nephews than vice-versa. But the asymmetry may confuse many users…

I am chuckling a bit as I read this. Although my situation may not apply to most others it could. I am related to my wife’s sister in law (she is only related by in law/ marriage via her brother). I am also related to her uncle by marriage (her blood aunt). This means our children have a double relationship to his grandchildren. Even if quite distant. One thought would be to double check the relationship calculator to make sure there is no other weird stuff.

The bottom line is that there is questionable things going on and without any info or guidance of what methods RM is using to make calculations we just guessing.

Here is two quick tests I did.
I exported a kinship report from my son-in-law POV (so list would be 900 something )

The kinship has 978 people but using a query the list had more people (will dig into that later) my guess it may have to do with people being married before and/or after (see snip it below)

Also I test in Family Historian and I believe # of people were identical (kinship FH to kinship RM). Not sure if RM is using shortest path or if any path is true etc.

In the case where you are missing the relationship, you need to check the set relationship/ relationship calculator in reverse as in my sis-in-law case, it doesn’t show a relationship one way BUT does the other…

Even in the wrong relationship such as Squire being a spouse of, when I reverse the set relationship/ relationship calculator it gives me the correct one…

and no it is NOT necessarily using the shortest path…

As for weird family relationship–we have a lot as Squire and Louisa Green had 16 other siblings–all except 1 lived long lives with many complicated relationships— on the Ketchum side we have a uncle/ niece who were also brother-in-law/ sister-in-law–another where Mom-in-law and dau-in-law were also sister-in-laws— but the best is hubby is 4th cousin to his sister and 4th COUSIN to himself!!!

that actually would explain some things – though I do not fully understand why/how that is.

and @thejerrybryan— having 2 spouses is NOT the problem as I have many who had 2 spouses that show up right as ’ spouse of’ for the blood relative and no relationship for the 2nd spouse of the non-blood relative…

and Tom the asymmetry is very confusing-- going back to Squire Green showing as spouse of instead of 3rd g-- well Squire was married twice-- so when I go to his 2nd wife ( with no kids) her relationship says
image

so IF I did NOT know what was going on here, I would be wondering which one of Squire’s 8 brothers she was married to!!!

Thanks for all the info
edit wouldn’t mind doing a small drag n drop or small gedcom BUT not a large one IF I MIGHT lose other info…