# How to tell whether Object is a duplicate?

Say I duplicate (not linked duplicate) the default Cube, "Cube", N times and name the duplicates "Cube.001" to "Cube.NNN". How can I tell, aside from the name, that "Cube.NNN" is a duplicate of "Cube"? Is it possible to determine the object, that "Cube.NNN" was duplicated from? I need this for my export program to avoid exporting the geometry of duplicates.

• The question doesn't if this is talking about object or mesh data. Mar 11 '16 at 9:02
• You have a point, but the question, as it is, refers to pure object duplicates, with meshes and all else. Mar 11 '16 at 9:17

This can only be done reliably by comparing all the mesh data (every vertex, weight, normal, UV ...etc).

The automatic name given to newly duplicated objects is always the original name plus some postfix (.001, .002 etc). As soon as you manually rename the duplicates then that 'thin' link to their past also goes. When you duplicate using Ctrl+D you also duplicate the Mesh.

You can Duplicate linked (Alt+D), which creates a new object, but references the mesh of the original object. Duplicate linked does let you see where it came from. object.data.name

Because I don't think you'll find any quick solution, for future projects you might consider a naming convention for your duplicated objects so that they do include the name of the original.

If you named your duplicates with the convention

Cube -> Cube0, Cube1, Cube2, Cube3, Cube4


Then sure, then you do

obj = bpy.data.objects['Cube0']
original = obj.name[:4]
mesh_to_export = bpy.data.meshes[original]


Naturally not all objects will be duplicated from a Cube, you could use regex pattern matching if the naming convention is consistent. Some people don't like regex, but whatever. I'm assuming you know how to write a regex if you are writing an exporter.

Or even a loop to test if the obj.name starts with any of the following ['Cube', 'Plane'] etc.

• I think OP is looking for a python solution.
– p2or
Apr 22 '15 at 8:22
• yeah, but I mention the keyboard shortcuts anyway. I don't think there is a python solution, not something that would work retroactively anyway cross sessions. Apr 22 '15 at 8:24
• @zeffii Thanks. Is the linked reference retained somewhere in the object? Apr 22 '15 at 8:32
• mmhhh, but I think it's possible to compare the geometry even if it would take long time for calculation.
– p2or
Apr 22 '15 at 8:33
• yeah @user1095108 the object.data.name will show the name of the mesh used. --- @poor, yeah that's always possible, I think there's even a Mesh diff tool which can compare meshes, never used it.. Apr 22 '15 at 8:35

It's possible to check the vertices of both meshes via unit_test_compare for a first quick test:

import bpy

first_obj = bpy.data.objects['Cube']
first_obj_mesh = first_obj.data

second_obj = bpy.data.objects['Suzanne']
second_obj_mesh = second_obj.data

print(first_obj_mesh.unit_test_compare(second_obj_mesh))


Console Output:

Number of verts don't match

For Linked Duplicates iterate through the mesh objects in the scene and check if mesh data block is associated with multiple objects:

import bpy

obj = bpy.data.objects['Cube']
obj_mesh = obj.data

mesh_dict = {}
for i in bpy.data.objects:
if i.type == 'MESH':
mesh_dict.setdefault(i.data, []).append(i)

print(mesh_dict[obj_mesh])


Note: Of course, it's always possible to compare the geometry, even if it would take long time for calculation and a bit of math, depending on your requirements.

• I get the following error when testing your code: KeyError: 'bpy_prop_collection[key]: key "Suzanne" not found' Oct 1 '18 at 14:37
• Add a monkey or any other geometry and rename it to Suzanne @HelloGoodbye
– p2or
Oct 13 '18 at 19:13

if object.data.users are bigger than 1, you already know that any of those meshes are the same than original so you can use any of them as "original".