Neo4j py2neo合并

Question

I have some questions on how to leverage graph.merge() in py2neo. I am trying to search in Neo4j to find if there is an existed account_node and link it with payment_change_event.

        account_node = graph.find_one('Payment_account', property_key="payment_account_id", property_value=payment_account_id)
        if not account_node:
            account_node = Node('Payment_account', payment_account_id=payment_account_id,\
                email=email, bank_name=bank_name, account_number=account_number
        changed_to_edge = Relationship(payment_change_event_node, 'CHANGED_TO', account_node)
        graph.create(changed_to_edge)

I am thinking of changing the logic to use graph.merge() like below instead of using find_one() and update the graph.

            account_node = Node('Payment_account', payment_account_id=new_payment_account_id)
            graph.merge(account_node)
            account_node['email'], account_node['bank_name'] = email, bank_name
            account_node['account_numner'] = account_number
            graph.push(account_node)
            changed_to_edge = Relationship(payment_change_event_node, 'CHANGED_TO', account_node)
            graph.create(changed_to_edge)

However, my question is by doing the above I will need to rewrite all the nodes all the time.

Is there any way for py2neo that can merge all the properties(even for non-constraint property)

Let's say account_node has payment_account_id as the constraint property and email, bank_name and account_number are non-constraint properties.

Now I already have had this account_node in the neo4j.

account_node: payment_account_id='123', email='abc@abc.com', bank_name='mybank', account_number='123456'

Then I want to run the process to change it to check if this node is changed. If some non-constraint properties changed I want update them.

If not I don't want to touch it but just link it to a payment_change_event_node.

I try to do

Node('Payment_account', payment_account_id=payment_account_id, email=email, bank_name=bank_name, account_number=account_number)

But it will complain on the merge().

File "test_graph.py", line 159, in main
    graph.merge(account_node)

py2neo.database.status.ConstraintError: Node(443971) already exists with label `Payment_account` and property `payment_account_id` = 'xxxxxxxxxxxxxxx'

Any efficient way to achieve that in py2neo? Is there any way like this? I hope there is a function only merges on constraint properties and update the changed non-constraint properties.

account_node = Node('Payment_account', payment_account_id=payment_account_id, email=email, bank_name=bank_name, account_number=account_number)
graph.merge(account_node)
graph.push(account_node)
changed_to_edge = Relationship(payment_change_event_node, 'CHANGED_TO', account_node)
graph.create(changed_to_edge)

Answer 1

When you MERGE a node neo4j looks for all properties of that node.

If it does not find a node that has all properties, it will create a new one.

In your case, a Payment_account node with the payment_account_id exists already. When you merge a node with the same payment_account_id but different email , bank_name properties, neo4j tries to create a new node with the specified properties.

Since you have a uniqueness constraint on payment_account_id , the node cannot be created.

The solution is to MERGE only on the unique property and then update the other properties (like you did in your example).

If you do not want to set node properties every time, you have to check if the property exists already.

account_node = Node('Payment_account', payment_account_id=new_payment_account_id)
graph.merge(account_node)
# check if you have to set a property
if account_node['email'] != email:
    account_node['email'] = email   
graph.push(account_node)