PyMongo: Removing a nested object without knowing the key - mongodb

Question

Let's say I have a collection that looks like this:
{
'_id': ObjectId('abc'),
'customer': 'bob',
'products': {
'1234':
{'name': 'Shirt',
'productID': 5
},
'5679': {
'name': 'Hat',
'productID': 5
}
}
'1011': {
'name': 'Jeans',
'productID': 9
}
}
}
I am looking to remove all nested objects whose 'productID' property is 5, so the collection would look like this afterwards:
{'_id': ObjectId('abc'),
'name': 'bob',
'products': {
'1011': {
'name': 'Jeans',
'productID': 9
}
}
}
I know the following information:
customer: bob
productID: 5
Is it possible to do a wildcard on 'products'? Something like this (it does not work):
db.update({'customer':'bob'}, {'$unset': {'products.*': {'productID': 9}})

Answer 1

If you have a choice, refactor your data to make each item a list element, e.g.
{
'customer': 'bob',
'products': [
{'code': '1234',
'name': 'Shirt',
'productID': 5
},
{'code': '5679',
'name': 'Hat',
'productID': 5
},
{'code': '1011',
'name': 'Jeans',
'productID': 9
}
]
}
Then your update becomes a piece of cake:
db.mycollection.update_one({'customer': 'bob'}, {'$pull': {'products': {'productID': 5}}})
result:
{
"customer": "bob",
"products": [
{
"code": "1011",
"name": "Jeans",
"productID": 9
}
]
}
Persisting with poor choices of schema will no yield long term rewards.