Ecommerce - MongoDB Applied Design Patterns - page 100

function passed to its constructor. In this case, we're using the list function to cre-
ate an empty list when the given parent isn't found.

Once we have this subgraph, we can update the nodes as follows:

def def update_node_and_descendants (

nodes_by_parent , node , parent ):

# Update node's ancestors
node [ 'ancestors' ] = parent . ancestors + [
{ '_id' : parent [ '_id' ],
'slug' : parent [ 'slug' ],
'name' : parent [ 'name' ]} ]
db . categories . update (
{ '_id' : node [ '_id' ]},
{ '$set' : {
'ancestors' : ancestors ,
'parent' : parent [ '_id' ] } })

# Recursively call children of 'node'
for for child iin nodes_by_parent [ node [ '_id' ]]:

update_node_and_descendants (
nodes_by_parent , child , node )

In order to ensure that the subgraph-building operation is fast, we'll need an index on the an-
cestors._id field:

>>> >>> db . categories . ensure_index ( 'ancestors._id' )

Rename a category
One final operation we'll explore with our category hierarchy is renaming a category. In order
to a rename a category, we'll need to both update the category itself and also update all its
descendants. Consider renaming “Bop” to “BeBop,” as in Figure 5-4 .