Received error in saving edits

What and who is affected by this issue/request?

Anyone trying to save and upload their edits when whatever triggers this error message is present in them

Where on the platform does it happen?

When you are saving your edits to upload them

2022-07-30T07:00:00Z## How do we replicate the issue?
I believe you would have to have an error in your edits to trigger this error message

Expected behavior (i.e. solution)

My edits should have been sent to the server

##error message:
Errors occurred while trying to save
Placeholder relation not found for reference -14 in relation -15

##help needed
I have no real idea of what this is about other than something to do with relations hence I have no clue as to how to fix it. In the meantime, I cannot edit because I have already received the “You have too many edits to back up” message. Your help would be greatly appreciated.strong text

Without any links to data that might be causing it, I’d be surprised if people could provide specific help.

General advice might be “save early, save often”.

If you can remember what you did, and can reproduce that on the dev server (search “sandbox for editing” in the OSM wiki), then it might be worth reporting that as a github issue against iD.


I have been wondering the same thing. There is a way to download the changeset that I am trying to save. Do you think that might help duplicate the problem and find a fix?


1 Like

If you can save your attempted changes as a.osc file which contains relations -14 and -15, and can shed some light on how those got created, it might be possible to understand what happened.

If you can save the .osc file then you might be able to manually remove some parts of it (with a text editor) then you might be able to upload that using JOSM.


I was able to do that but it is pretty much Greek to me. There are 5 instances of the word “relation” in the file but still don’t see what the editor program’s problem is.


(Attachment changes.osc is missing)

If you can upload it somewhere someone might be able to have a look. Probably not me (on a phone), but someone might.

To clarify: the error is directly produced by the OSM editing API and indicates that the new relation with id -15 has a reference to the new relation with id -14 that however isn’t in the uploaded data. Negative ids are used in editors as placeholders that are replaced by new positive ids when you upload. Obviously this is likely an error in the editor that needs to be investigated.

It would be really helpful if you could attach the .osc file, you’ll probably have to change the extension to .txt for that to work.

1 Like

Another possible explanation could be that relation id -14 in being created only after relation -15, see for discussion.

Without actual .osc payload data, it’s impossible to tell what’s the issue here.

The related iD issue was closed 5 years ago so it can’t really just be ordering on its own. Could naturally be a cyclic reference but iD is supposed to complain about those too.

Well, I’d say it’s still fairly trivial to recreate the issue using circular dependencies. Add a few new relations, add some cross-references, upload, and…


<osmChange version="0.6" generator="iD">
<node id="-1" lon="-63.61725917132719" lat="44.64067645151963" version="0">
<tag k="name" v="test"/>
<relation id="-3" version="0">
<member type="relation" role="" ref="-2"/>
<tag k="type" v="enforcement"/>
<tag k="enforcement" v="maxspeed"/>
<tag k="maxspeed" v="100"/>
<relation id="-2" version="0">
<member type="node" role="" ref="-1"/>
<member type="relation" role="" ref="-3"/>
<tag k="type" v="site"/>
<delete if-unused="true"/>
1 Like

With help of several people on this forum this problem was finally bypassed. That only led to a new error having to do with either malformed or missing redirect url. More editing finally made that one go away, too. Thanks to all of you for your help. I could not keep editing if I could not upload what was already in the changeset because tables would start to overflow when there more edits than the system can handle. It was a case of being between a rock and a hard place, With your kind and useful replies, it was all taken care of.


1 Like

Could you be a bit more specific, maybe, as it’s incredibly difficult to find out what this error is about?