Bug related to Node/Edge tables?

guruper2

After initial commit, I still get all Node and Edge tables in the database in the list of changes. The compare engine says it's a change in the Primary key. When trying to commit the change, I get a git error, probably because there's no real change in the underlying files. Is there a workaround, or do we need to wait for a bugfix?

DanC

Hi @guruper2

Can I ask what the Git error message you're getting is?

guruper2

The error is: Failed to execute git command 'commit --allow-empty-message -m x'

I think the underlying issue is that there are no actual changes to check in. The local files are correct, but the UX comes up with "fake" changes when refreshing the commit window.

DanC

Hi @guruper2

The issue Git is having is you're leaving the commit message empty and I assume you have it configured to enforce a message. 

guruper2

The git command is the same for all my database projects and the comment i posted was "X". So that's not the issue here. The real problem is that the script being generated is wrong. All the files checked in to GIT is OK, but something special is happening to Node and Edge tables because they have "hidden" fields, and they are not scripted. Even though the $edge_id and $node_id fields are hidden in the table script, you can still put primary keys on them. The script should be: ALTER TABLE [Node].[Address] ADD CONSTRAINT [PK_Node_Address] PRIMARY KEY CLUSTERED ($node_id) ON [PRIMARY]

But it shows as: ALTER TABLE [Node].[Address] ADD CONSTRAINT [PK_Node_Address] PRIMARY KEY CLUSTERED ON [PRIMARY]

This causes all node and edge tables to show up having a difference, but the differences are not real. It's because the primary key is scripted wrong. 

DanC

Hi @guruper2

In that case I'll need more information from yourself and so I'll reach out via a support ticket to yourself

guruper2

Thanks