Retrieving Nodes

Out of the box, the Maya API is a bit cumbersome when it comes to retrieving DG and DAG nodes from a scene. This usually leads each TD to write their own code for the task, and this is also something that Bana aims to provide.

The goal here is to offer a higher-level set of methods allowing to retrieve nodes with enough flexibility to cover most of a TD’s needs while remaining as fast as possible.

Since these methods are in fact iterators, it is easy to build on top of them in the case where more filtering options are required, such as for example skipping the DAG shapes that are templated.

But what sets this library apart from the usual implementations is its well-defined pattern matching specification. When Maya’s interpretation of the wildcard character * is everyone’s guess, Bana offers both precise and predictable results.


In Bana, there are 2 groups of classes from where the scene nodes can be retrieved:

The former group represents the most common use case while the second can be used to slightly speed things up when the extra functionalities brought by the MFn* classes are not required.

For each of these classes, two types of methods are then exposed as the API:

  • the methods starting with bnFind which return an iterator over a collection of nodes matching the input filters.
  • the methods starting with bnGet which return a single object matching the input filters. If zero or more nodes are found, then None is returned.

When using the lower-level family of classes, it is possible to explictely pass a node type to match through the fnType parameter but in the case of the function set classes, no fnType parameter is defined. Instead the node type to match is deduced from the calling class. In other words, a call to maya.OpenMaya.MFnDagNode.bnFind() will match any node of type maya.OpenMaya.kDagNode while calling maya.OpenMaya.MFnTransform.bnFind() will only match transform nodes.


Node names are available from the maya.OpenMaya.MFnDependencyNode class but not directly from the maya.OpenMaya.MObject one. As a result, retrieving maya.OpenMaya.MObject objects from a pattern will internally convert them to maya.OpenMaya.MFnDependencyNode, in which case there won’t be much benefits from using the MObject.bnFind() method in place of MFnDependencyNode.bnFind().

DG vs DAG Nodes

Although it is possible to iterate over all DG or DAG nodes using the methods exposed within the MFnDependencyNode and MObject classes, it is not possible to filter DAG nodes this way using a path pattern. Indeed, these methods only accept name patterns.

Hence it is recommended to use instead the methods defined in the classes MFnDagNode and MDagPath whenever DAG nodes are to be retrieved. Furthermore, these offer a boost in performances, especially when only a specific branch of DAG nodes needs to be traversed through the use of the bnFindChildren() and bnGetChild() methods.


>>> import bana
>>> bana.initialize()
>>> from maya import OpenMaya
>>> # Retrieve a transform node named 'root'.
>>> root = OpenMaya.MFnTransform.bnGet(pattern='*|root')
>>> # Recursively iterate over all the DAG nodes child of 'root'.
>>> for node in root.bnFindChildren():
...     print(node)
>>> # Find all the mesh nodes in the scene containing the word 'Shape' but
... # not belonging to any namespace.
>>> for node in OpenMaya.MFnMesh.bnFind(pattern='*|*Shape*'):
...     print(node)