Can't delete unlinked component

[SGoldthwaite]

I can't delete a component from my project that I inserted into my design.  I unlinked it, but still can't delete it .  See screencast

http://autode.sk/1fLJzD4

[TrippyLighting]

I'd try saving and then closing the current design you are working on and then delete the part.

[SGoldthwaite]

---

@TrippyLighting wrote:

I'd try saving and then closing the current design you are working on and then delete the part.

---

I tried that, but it didn't work.  What did work is I saved my Fusion design to a local file to my hard drive, then deleted my design from the fusion project, then I was able to delete the part from my project.

[TrippyLighting]

You should not have to do that. This sounds like a nunecessary limitation or bug!

Breaking the link should be all you need to do.

[jeff_strater]

The reasons why this works this way are twofold:

1. We made a design decision in Fusion that we would not allow missing references to happen.  Allowing missing references causes lots of problems, both for you (users) and us (developers).  So, we try very hard to prevent you from deleting external designs that are being referenced.  That may change at some point in the future, but for now, for better or worse, that's what we have.
2. The way that version management works.  Say you have design A that references design B.  Both are at version 1 (A(v1) and B(v1)).  Then, you edit A, and delete the reference to B.  Then, you save A.  So, now you have A(v2), and B(v1).  So, the current version of A does not have a reference to B.  But, A(v1) still exists in the data management system, and you can "promote" A(v1) to become A(v3), and all of a sudden, you have a reference to B again.  So, when we check whether B can be deleted, we have to go back through all versions of A to check.

So that's the "why".  I understand and agree that the workflow described here is confusing and frustrating.  There are two possible solutions to this:  Either Fusion allows you to purge back versions of a design (which is essentially what @SGoldthwaite did by saving local and re-uploading, if you think about it), or we allow Fusion to tolerate missing references.  Both are valid requests, IMO.  They probably should be on the Idea Station.

Hope this helps clear up the mystery

Jeff Strater (Fusion development)

[SGoldthwaite]

I feel like I'm trying to follow the timeline in a time travel movie:)  I didn't really follow that, but I see your point.  Earlier in my design I inserted a ball plunger from McMaster-Carr and then broke the link and I was able to delete the ball plunger from my project.  So, when can I break the link and delete the original and when can I break the link and still not delete the original?  Is there a way to insert a component without in linking back to the original?  I think it would solve a lot of problems for things like this.  Small hardware like the things inserted from McMaster are unlikely going to be modified.  I'd prefer to insert it without any external reference.

[jeff_strater]

you mean like  Primer?  (I love that movie, but I had to watch it 4 times to even begin to understand it).

It all depends on whether you have saved a version with the reference or not.  If you Insert an external component, then immediately do a Break Link on it (without saving), then that link is never recorded, so you can delete the referenced design.  But, once you save with that link, you're stuck.  The best thing at that point (and I admit that this is lame) is to create a "junk" folder in your project, and move the designs you are no longer interested in into that folder.  Fusion will track references across folder moves.

None of this is impossible to fix in future Fusion versions, it's just a question of priorities.

Jeff

[TrippyLighting]

Perhaps the simplest implementation would be simply to insert the external file that way it was done before without any ties. To be offer as a second alternate  option to inserting an X-Ref . 

Who would want to keep track of al the nuts bolts and myriad of parts that get added to a design that are simply downloaded from McMaster Carr, Misumi and Another myriad of other vendors. A lot of these parts are never edited, only exist in version one and keeping track of versions is really not necessary.

[jeff_strater]

Yeah, @TrippyLighting, I agree.  We discussed that (having two insert commands:  Insert and Insert XRef), but it was decided that this would cause confusion ("which Insert should I use?").  I can see that argument.  I don't particularly agree with it, but I can appreciate the point.

Jeff

[SGoldthwaite]

---

@jeff_strater wrote:

Yeah, @TrippyLighting, I agree.  We discussed that (having two insert commands:  Insert and Insert XRef), but it was decided that this would cause confusion ("which Insert should I use?").  I can see that argument.  I don't particularly agree with it, but I can appreciate the point.

Jeff

---

How about making it an option in the preferences?  For user's that wanted both insert methods, they could enable that option.  This way it's  likely that the users who turned this on would understand how it works.

[pdevDMGRW]

3.5 YEARS LATER - and this f**** bug (yes, bug) seems to be even worse.... because now, I can't even use @SGoldthwaite 's workaround (saving the file local, and deleting from the cloud) cuz I'm prohibited from saving local if there are *any* linked components whatsoever (*any* linked components, not just the busted/broken remnants we are trying to purge - ie. the purpose  of this whole thread).   Error is "Can't save .f3z for blah blah...linked components".   So I have to break link ALL components, then it'll save local.  But I can't delete the cloud version because it is referenced.... *somewhere*.    And re-loading the local file (after it has been renamed in ms-dos, of course, just in case)...  I can't hard delete the busted obsolete components in question.  All for naught.

Anyone know what I'm doing wrong????      If the busted component is part of some larger component you use often, the busted crud will cascade through and pollute EVERY design you do, FOREVER.    

Jeff the Autodesk guy said: "we try very hard to prevent you from deleting external designs that are being referenced".    Dude... please.... I'm begging you...   don't try so hard.

Bah humbug.

[TrippyLighting]

I've had no problems saving assemblies locally. I usually just use the public link in the data panel and download through the web interface.

I've done this many times and I donut remember it ever failing.

[Anonymous]

If logic can't be changed after so many request for years.  At least, it would be nice to be able to find the part we want to delete in every assemblies it is referenced (and also in what version) it is used.  F360 should know where the part is located because it prohibe us to delete it.  Actually, the part is not referenced because it is not located in the last version.  With that system it seems impossible to ever delete the bad part when it has been used in so many assemblies.

thanks

[ipmcc]

I found this thread while googling for "fusion cant delete or remove component".

This appears to still be an issue in 2025. It's unbelievably frustrating. How is a person supposed to create long-running, well-factored, clean designs if they can never 'reverse' an action once it crosses file boundaries and is saved? This is nuts. I hear you; I understand why it's hard, and I see the failure modes you're trying to avoid. But it's really a big roadblock to creating larger designs in Fusion. Here's an example:

• My goal is to design a "motor mount" for, say, a stepper motor, that's part of a larger design. Call it MM1.
• To get the bolt-hole spacing right, I import a STEP file of the motor I got from the manufacturer, call it SM1. OK, looks good.
• Save MM1 (which now has SM1 in it.)
• Create/Open "larger assembly" file, call it LA1. 
• 'Insert Derived' and choose the motor mount from MM1 (of which SM1 is a sub-component).
• Work, save, work, save, work, save, and so on.
• Create "even larger assembly" file, call it LA2.
• 'Insert Derived' and choose LA1 (of which SM1 is a sub-component).
  • Now, I've got four conceptual levels: SM1 > MM1 > LA1 > LA2 but I've also got three instances of SM1 (the original, the version imported to MM1, and the version imported to LA1, each at a different level in the inheritance tree, but each is referenced by another file.
• Work, save, work, save, work, save, and so on.
• Realize that now you have 2 copies of SM1, because before you had done MM1, you had ALSO imported SM1 into LA1.
• The excess instances of SM1 can apparently never be deleted.
• The excess instances of SM1 can't be "Removed" either.
• The excess instances of SM1 can't even be hidden. If you make them invisible, then focus and select the owning component, there are ghosted selection indicator UI for the part that's hidden. 

As I get further down the road with Fusion, it feels like long-lived designs that span multiple files will almost inevitably get cruft-ed up with references to old crap. So, while I accept the argument about not allowing deletion of referenced components, it's going to be extremely painful in any sufficiently large design that spans multiple files. It's clear that the intent is for people to use multiple files in complex designs -- timelines quickly become quite untenable if you try to work on even a dozen (or so) components in one file -- but the idea that you can never correct a mistake that happened before your last save? That just means that you effectively can't have complex, long-lived designs. Software used to encourage that you "save early, save often". It's nice that Fusion doesn't crash ALL that often, but the idea that "Saving" a file is actually, functionally destructive? That's not intuitive at all! Saving should be about security, not about stripping away capabilities.

I suspect the answer here will be, "If you need that kind of scalability, you probably need to use a different CAD package." But the idea that you can never correct a mistake if it's ever been derived into another file, that has then been saved? That just feels bonkers to me. Also, none of the designs I'm working on have multiple people working on them -- it's just me here, and I'm getting burned by this, over and over.

There are lots of great examples of (document) models that are resilient against issues like this: Git, BitCoin, etc. They remember what everything looked like at every stage and can pull data from history, at will. I can understand the desire to not re-invent those sorts of wheels here, but 'you can never delete anything that's been referenced'? That seems like a major issue if you want customers to be able to make complex designs without having to start over every time the weight of their prior mistakes becomes onerous... 

I long ago accepted that I would have to recreate designs a second time once I fully understood the design in its final state, but to be unable to delete a vestigial component? It's really quite an impediment to coming out the other side with a well-factored, survivable design. I'm sure the Autodesk engineers understand this; Software Engineers don't make large commercial apps out of one giant source-code file. It would be impossible to collaborate! But this approach also makes it possible to delete cruft at some point. The document management for fusion appears to simultaneously encourage a separation of concerns while also making it impossible to maintain a cluster of concerns in an ongoing endeavor.

This is extremely frustrating. I won't hold my breath, but this feels like a major handicap for Fusion.

-Ian

[TrippyLighting]

I thought I'd mention that Insert Drive is not the same as a linked component.

I believe the newest update addresses the request in this thread. If you read the "what's new", which is accessible from the help menu in Fusion:

There is a section called "Trash while referenced".

[ipmcc]

Thanks! That sounds like a solid step in the right direction, but it's really not the same as the issue that I described. I want to delete one or more components within a design, but not the entire design. In short, I want to be able to treat it like source control (i.e. Git). It's good to see that there has been movement in that direction.

- Ian