GEDCOM Fix

"Some Records Were Not Uploaded Correctly" — What This Actually Means

FamilySearch's vague "some records were not uploaded correctly" message explained: what's likely missing, why it happens, and how to check what you lost.

GEDCOM Fix Editorial TeamUpdated 3 min read

Of all the vague states a FamilySearch GEDCOM upload can end in, "some records were not uploaded correctly" is the most specific-sounding — and still tells you almost nothing about what actually happened. This guide covers what the message is likely referring to and how to find out for yourself.

What we know about when this appears

This exact message has been reported by FamilySearch Community users on files in the range of roughly 1,000 individuals — large enough to strain the "compare and add" process FamilySearch uses to match your file's people against what's already in Family Tree, but not so large that the upload fails outright. It sits in the middle of the failure spectrum: not a hard rejection (that happens closer to the documented 100 MB file-size cap), and not the total data loss reported on some much larger files, but a partial, unspecified gap.

What's most likely happening

FamilySearch's import process isn't a raw file copy — it runs each individual and family record through a comparison against Family Tree's existing data, looking for matches, potential duplicates, and conflicts. At scale, this comparison step is where records most plausibly get dropped, skipped, or silently merged into an existing record in a way that doesn't preserve everything you sent. The message is generic enough that it could also point to a handful of records the importer simply couldn't parse — a malformed date, an unexpected structure, an encoding issue in a specific name — without flagging which ones.

Because FamilySearch doesn't publish the internals of this process, we can't tell you definitively which of these is happening in your specific case. What we can tell you, based on the pattern of reports, is that it correlates with individual count and file complexity generally, not with one particular record type.

How this differs from total silent data loss

It's worth distinguishing this message from the more severe pattern where a file uploads with no error at all and the vast majority of people are simply missing afterward — one reported case involved a file with over 2,500 individuals resulting in only two people appearing in Family Tree. "Some records were not uploaded correctly" is FamilySearch actually surfacing that a problem occurred, even without details; total silent loss gives you no signal whatsoever. If you're seeing this message, treat it as the system telling you something is wrong, not as evidence your file is fine — but it's a meaningfully less severe outcome than the alternative.

How to find out what you actually lost

Since the message itself gives you nothing to go on, the practical approach is comparison: note your file's individual and family counts before uploading (any GEDCOM-aware tool, including the Upload Readiness Checker, will show you this instantly), then check the resulting count in your FamilySearch Family Tree afterward. A meaningful gap between the two tells you records were genuinely dropped, not just merged into existing matches — which at least turns a vague warning into a concrete number you can act on.

What to do about it

The most consistently reported fix — echoed across multiple FamilySearch Community threads — is the same one that fixes the size-based failures: split the file into smaller pieces before uploading, so each piece has fewer records for the compare-and-add process to work through at once. See How to Split a Large Family Tree Before Uploading for how to do that without breaking the structure of the file you're left with, and run the result through the Upload Readiness Checker first so you're not guessing at whether the smaller file is actually a safer bet.

Do it now

Keep reading

Common questions

Does this message tell me which records were affected?
No. It's a generic notice with no list of what was skipped, which is why comparing individual counts before and after is the only reliable way to know what actually happened.
Is this the same as an outright upload failure?
No — an outright failure means the file was rejected and nothing was added. This message means the upload technically completed, but part of what you sent didn't come through, which can be easy to miss since there's no error state.
How do I avoid this message happening again?
Splitting the file into smaller pieces before uploading is the most consistently reported fix, since this message correlates most with individual count and file complexity, not any single record type.