A few hundred curves result in huge file size

This is super strange… I have around 1000 polylines in this file, which consist each out of 4-10 line segments, and the file size is 5 GB.

I checked if it could be anything else, purged the file, but it seems to be this relatively dense chunk of polylines. When I export them, the exported file is equally large.

Anyone had this before?

Hello- You have a lot of block definitions - 400+. Did your Purge include unused block definitions? Please run Audit3dmFile and post the results…

-Pascal

this is super weird… I have purged and deleted all blocks, there are no blocks in the manager.

Even the audit size is too large to copy/paste, so I attached a txt file.

audit.txt (165.5 KB)

Hello - there are a lot of these:

7565402 bytes, id = 027e18ea-df92-4e2f-97e5-d7677f7bcb5f, name = "Curve: PolylineCurve m_pline[] is not valid."

That would do it. I recommend selecting the objects you care about and running Invert - is anything selected? If so, Delete. If not, Export to a new file…
Question is, how did this happen? if you run SelBadObjects and find anything, please send us one of those.

-Pascal

Hi Pascal,

There are almost no bad objects. The red polylines were baked from GH but something must have gone terribly wrong. If I explode them the file becomes even bigger although they are just single lines?

In GH, I made a convex hull around a hand full of points. They resulted in Polyline Curves, and I only baked those that were valid.

I have uploaded the whole file here if you want to have a look. I might have to look more into the GH process to figure out what went wrong.

Download (5gb): https://fil.email/HhESCVqF

I guess the extra is in object attributes - user data, user text or something that … that is attached to every object… I’ll take a look.

-Pascal

I think you’re right. There seem to be key values with huge amount of text. Where would they have come from??

Yup, that was it. If I selected one item and switched to view the user values Rhino would get stuck trying to show what looks like a LOT of text. I was able to select them all, in which case the text wasn’t shown, and I deleted all values. Back to 8mb.

Curious how those were baked from Rhino.

Thanks, Pascal.

It is ‘Grasshopper bake Data’ It would be good if you can reproduce this… it looks like the data is the GH archive itself.

@DavidRutten - in case this is of interest…

-Pascal

Grasshopper only includes data on baked objects when they are baked using the Geometry Cache object. Regular bakes do not add any user texts. However it is possible that objects are baked via scripts or plugins in which case the behaviour might change.

So yes, knowing how this happened would be of interest.

Hi David,

This is very strange but then again I used a really old GH definition that I changed around a bit to get the result. That shouldn’t have had any effect on the geometry though:

I used collections of points to make convex hulls, then I used the group component twice to have nested groups one to group each branch, the other one flattened the whole data), then I baked the last group component using the right click on the canvas.

Getting nested groups didn’t really work when I checked the baked geometry it was just one group) but I didn’t really care and kept baking multiple instances of it.

Hi Oliver -

There’s not enough information in this last post of yours for us to be able to determine how user text was added. Are you able to post a Grasshopper definition that lets us reproduce that behavior?
-wim