Jump to content


GlenB's Content

There have been 10 items by GlenB (Search limited from 02-April 19)


By content type

See this member's

Sort by                Order  

#96798 Why does FS think there is new information?

Posted by GlenB on 17 March 2020 - 02:44 PM in FamilySearch Family Tree

Thanks, I think I get it. Nasty of them to create so much work for us!

 

If I understand, you are comparing PersonTable.EditDate with LinkTable.extDate. What's the LinkTable.Modified flag for? And EditDate must be when I edited the record, not a date where I merely looked at in the FS Central comparison.

 

I have just over 16500 records in LinkTable and they ALL have the extDate = 0.0. Only 1100 of those have Modified = 0.

And I have about 17800 records in PersonTable and the EditDates are all > 0.0. Meaning in part that all but about 1300 of my people are matched up.

 

So theoretically, every single person who is linked to a FSFT ID should be in the list to be checked in Family Search Central. But I just processed a few hundred more yesterday and they are no longer in the list. Is that what the Modified does? The ones still to be checked are those with Modified = 1 which is about 14400 records?

 

And if I reset the count through the GUI does that make all the Modified = 0? At least until such time as I give Central the opportunity to go scanning and comparing the date fields.

 

I ask because I don't really want to go messing with the tables unless I know what I'm doing <grin> and at the moment I appear to have a TON of work to do!




#96763 Why does FS think there is new information?

Posted by GlenB on 13 March 2020 - 05:34 PM in FamilySearch Family Tree

But so many people? I have only processed about 100 people on the list so far, but only 5-10 of them had ANYTHING new.

 

Can you explain the algorithm that RM uses to decide if a person has new information or not?




#96760 Why does FS think there is new information?

Posted by GlenB on 13 March 2020 - 01:56 PM in FamilySearch Family Tree

I left the Family Search Central panel open for a while and it went and collected the list of my people who are linked to Family Search and for whom it thinks there is new information. Then I go look at that list to see what info I want to update / correct.

 

But here's the thing. MANY of those people in the list have nothing new about them, nothing at all. A simple example had a person with only a name, gender, birth date and parents. In the screen where my person's info and the FamilySearch person's info are compared, everything is green meaning (as far as I know) that the information is identical. Maybe it's the sources that are new? Nope, the tab at the top says Sources(0) so there are no Sources in FS that could have changed.

 

It's not just one person. Happens to me A LOT. and that wastes a lot of time Why does this happen?




#93811 Odd citation when importing a source

Posted by GlenB on 06 May 2019 - 05:07 PM in FamilySearch Family Tree

Sometimes I see something like this:

 

RootsMagic Sources

  Source

□ ◘ Ane Fredericksen in household of Niels Christensen, "Denmark Census of 1860"

      Citation: Census

      Citation: Census

□ ◘ .... more sources and citations

 

and sometimes something like this:

RootsMagic Sources

  Source

□ ◘ Ane Fredericksen in household of Niels Christensen, "Denmark Census of 1860"

      Citation: Census

      Citation: 

□ ◘ .... more sources and citations

 

But in all cases it goes away on the next refresh. These are not limited to Census type records, but happen on a variety of event types.




#93810 Sometimes RM thinks there are no FS Sources

Posted by GlenB on 06 May 2019 - 04:48 PM in FamilySearch Family Tree

A refresh would be helpful for sure.

 

I'm not suggesting RM should detect that FS isn't sending data. I'm suggesting that RM could notice when FS sends a DIFFERENT amount of data to a repeated request. If a particular FSID gives you 22 sources on one request and the next time you make that request you get 0 sources then a refresh sounds like a good idea. No matter what comes in the next time, you have to assume it is correct and carry on. If 22 comes in then thank heavens you refreshed and cleanup whatever was confused - and the user never even knew it happened.




#93809 Odd citation when importing a source

Posted by GlenB on 06 May 2019 - 04:43 PM in FamilySearch Family Tree

Agreed. I do run the DB tools fairly regularly. But, as I said, the spurious citation disappeared on the next refresh, not after reindexing. So perhaps the issue wasn't an index side-effect?

 

This supports my contention that the FSPT screen gets "confused" sometimes. There are several threads in this Forum where I point to behaviours that all suggest that might be what is happening. Situations where the page shows one thing and then after a refresh / redisplay shows another. Happens on the left side and on the right side. All of it seems related to counting how many lines of info to show under a particular heading. My years of programming suggest that "off by one" errors are among the most common yet hard to detect programming faults.

 

You know I would present you with a detailed "how to recreate" if I could. But the fact that I cannot reproduce it reliably doesn't mean it doesn't happen, and it happens often enough that it bothers me enough to write it up here.




#93805 Odd citation when importing a source

Posted by GlenB on 06 May 2019 - 03:12 PM in FamilySearch Family Tree

Done.

 

Was that meant just to make sure nothing had got messed up in the database. All seemed to be OK, all the DB tools were happy.

 

I fairly often see "spurious" citations appear under a source in RM when I'm importing FS sources. They always go away on the next refresh/redisplay. I only reported this one because it was the first time I had seen some other RM person referenced in one of those citations.




#93804 Sometimes RM thinks there are no FS Sources

Posted by GlenB on 06 May 2019 - 03:04 PM in FamilySearch Family Tree

Thanks, I didn't realize that refresh button had been put in the enhancement list. It would certainly help!

 

I still think RM is mis-counting, or maybe mis-interpreting what FS says, or stepping on its own data at some point. But even if none of that is true, RM is at least not detecting that something untoward has happened and refreshing on its own. Still, I suppose leaving it for the user to detect a possible error and giving them a way to force the refresh at least slaps a good bandage on the wound.

 

I would have thought that if the number of Sources changes from one display of the data to the next (ie you showed Sources(22) last time and Sources(0) this time) then it's a fair bet that something got messed up and perhaps the refresh in that case could be automated before the Sources(0) is even presented to the user.




#93800 Odd citation when importing a source

Posted by GlenB on 06 May 2019 - 02:00 PM in FamilySearch Family Tree

I was importing a source into RM from FS for a person using the FSPT screen, Sources tab.

After importing the Source the sides of the screen went blank briefly and redisplayed, as expected.

But on the left side (RM side) there was an odd citation shown. I'll type it here instead of trying to insert a screen clip in this message.

 

RootsMagic Sources

  Source

□ ◘ Ane Fredericksen in household of Niels Christensen, "Denmark Census of 1860"

      Citation: Census

      Citation: Susan Smith-1

□ ◘ .... more sources and citations

 

The issue is the citation to Susan Smith-1. Yes, this is the root person of the database, but Ane Fredericksen is her 4th cousin 4 times removed and was not explicitly referenced at any point since the source import was started. I had just imported that 1860 Census source from FS and attached it to a Census fact in RM.

 

I then imported some other source from FS to RM and, when the screen redisplayed, this odd citation was gone. And in checking, there is nothing in Susan Smith that references this source, so far as I can see.

 

So nothing got stored where it shouldn't have been, the data is OK, but the logic behind the display in FSPT is making mistakes - not consistently, so I realize how hard that is to locate!

 




#93799 Sometimes RM thinks there are no FS Sources

Posted by GlenB on 06 May 2019 - 01:16 PM in FamilySearch Family Tree

Can we try this one again please Renee? I think I got too complicated in my post, made too many suggestions, you focused on one and I think we didn't read the same thing on that one. So .... reset. I apologize for being long-winded, but I hope it helps.

 

Currently running version 7.5.9.0

 

This happens several times during an hour of updating RM from FS using the FSPT screen for people with changes.

 

I am updating RM by adding a few facts from FS and a bunch of sources as well. I add the facts first, then I go to Sources to transfer over the ones I want. Let's say there are 22 Sources in FS when I start, but none yet in RM, or maybe only a few. I follow the normal clicking to scroll down and select a Source and apply it to an RM fact. Then the left side and right side of the screen refresh - I don't know if they reload, but they go blank for a moment and then the lists redisplay. That's the normal process and it repeats until I'm done. And it all gets done eventually in the majority of cases.

 

But in several cases an hour the sides of the screen go blank and never redisplay anything. I wait, I'm patient, but nothing. So I go back to the Share Data tab which populates as expected. Usually I Edit RM Person to see if that helps. Nothing seems to bring back the Sources which were there moments ago. At some point in all of this I notice that the Sources tab is now labelled Sources(0) which surely is not correct, but it would explain why no Sources appear on the tab! And I think it gets stuck there. If I ignore this and move on to the next person they start out with Sources(0).

 

If I leave FSPT and open up the person in the regular RM views, click on the FS icon to the right of the name, it displays the Share Data tab fully populated. I go to the Sources tab, which is now labeled Sources(22), and finish transferring over the Sources I wanted. So the Sources are there and FS is providing them but for some reason, RM did not see them or count them and acts as if they are not there.

 

At the risk of getting too complicated, I will continue, because I think it might help describe the problem.

 

Sometimes in FSPT looking at the people who have new information I select a person and the Share Data tab populates as expected. I look at the data and everything I have in RM agrees with what is in FS, all green, so look at the Sources tab and it is labeled Sources(0). I click on the Sources and, yup, nothing there on the FS side. So I wonder why RM thought there was new information?

 

Then, looking at the facts in the Share Data tab I see some pretty detailed facts - full dates and locations for more than just BMD. It occurs to me that there probably are sources for these facts that are just not being shown to me. I go back to the RM screens and click the FS icon and there it says Sources(8), showing me the 8 Sources which were not presented in the FSPT version of this same screen.

 

The commonality of these 2 scenarios is that the FSPT screens act as though there are no Sources when in fact there are. Whether it is RM getting lost or FS speaking in tongues is a detail for you to figure out. But there clearly IS some way for incorrect data to be presented and for the user to be unaware of that. Absent the discovery of a counting error or data over-write or misinterpreted status message, might I suggest a bit of a brute force way to at least detect the problem?

 

What if, when a person gets loaded into that FSPT screen you track the FSID and the # of sources off to the side somewhere. If you ever requery FS for sources (as it appears that you must be doing because if not you are simply zeroing the counter by mistake) and if the FSID matches but the # of sources does NOT, then you know something happened. Logging this could lead to understanding if it happens very often. At that point, flush all buffers and restart as if the person was just being selected for the first time from the list at the left and hope that clears it up. This might not be sufficient to solve the problem, so maybe it's just instrumenting the code to make diagnosis more effective.

 

Willing to discuss if this still seems not clear.