base class — bpy_struct

class bpy.types.BlendDataLibraries(bpy_struct)

Collection of libraries




value (boolean) – Value

remove(library, do_unlink=True, do_id_user=True, do_ui_user=True)

Remove a library from the current blendfile

  • library (Library, (never None)) – Library to remove

  • do_unlink (boolean, (optional)) – Unlink all usages of this library before deleting it

  • do_id_user (boolean, (optional)) – Decrement user counter of all datablocks used by this library

  • do_ui_user (boolean, (optional)) – Make sure interface does not reference this library

classmethod bl_rna_get_subclass(id, default=None)

id (string) – The RNA type identifier.


The RNA type or default when not found.

Return type

bpy.types.Struct subclass

classmethod bl_rna_get_subclass_py(id, default=None)

id (string) – The RNA type identifier.


The class or default when not found.

Return type


load(filepath, link=False, relative=False, assets_only=False)

Returns a context manager which exposes 2 library objects on entering. Each object has attributes matching which are lists of strings to be linked.

  • filepath (string) – The path to a blend file.

  • link (bool) – When False reference to the original file is lost.

  • relative (bool) – When True the path is stored relative to the open blend file.

  • assets_only (bool) – If True, only list data-blocks marked as assets.

import bpy

filepath = "//link_library.blend"

# load a single scene we know the name of.
with as (data_from, data_to):
    data_to.scenes = ["Scene"]

# load all meshes
with as (data_from, data_to):
    data_to.meshes = data_from.meshes

# link all objects starting with 'A'
with, link=True) as (data_from, data_to):
    data_to.objects = [name for name in data_from.objects if name.startswith("A")]

# append everything
with as (data_from, data_to):
    for attr in dir(data_to):
        setattr(data_to, attr, getattr(data_from, attr))

# the loaded objects can be accessed from 'data_to' outside of the context
# since loading the data replaces the strings for the datablocks or None
# if the datablock could not be loaded.
with as (data_from, data_to):
    data_to.meshes = data_from.meshes
# now operate directly on the loaded data
for mesh in data_to.meshes:
    if mesh is not None:
write(filepath, datablocks, path_remap=False, fake_user=False, compress=False)

Write data-blocks into a blend file.


Indirectly referenced data-blocks will be expanded and written too.

  • filepath (string) – The path to write the blend-file.

  • datablocks (set) – set of data-blocks (bpy.types.ID instances).

  • path_remap (string) –

    Optionally remap paths when writing the file:

    • NONE No path manipulation (default).

    • RELATIVE Remap paths that are already relative to the new location.

    • RELATIVE_ALL Remap all paths to be relative to the new location.

    • ABSOLUTE Make all paths absolute on writing.

  • fake_user (bool) – When True, data-blocks will be written with fake-user flag enabled.

  • compress (bool) – When True, write a compressed blend file.

import bpy

filepath = "//new_library.blend"

# write selected objects and their data to a blend file
data_blocks = set(bpy.context.selected_objects), data_blocks)

# write all meshes starting with a capital letter and
# set them with fake-user enabled so they aren't lost on re-saving
data_blocks = {mesh for mesh in if[:1].isupper()}, data_blocks, fake_user=True)

# write all materials, textures and node groups to a library
data_blocks = {*, *, *}, data_blocks)

Inherited Properties

Inherited Functions