aarondl
commented
6 years ago "f_keys": [
{
"table": "account",
"name": "account_id_fkey",
"column": "id",
"nullable": false,
"unique": true,
"foreign_table": "entity",
"foreign_column": "id",
"foreign_column_nullable": false,
"foreign_column_unique": true
},
{
"table": "account",
"name": "account_id_fkey",
"column": "id",
"nullable": false,
"unique": true,
"foreign_table": "entity",
"foreign_column": "type",
"foreign_column_nullable": false,
"foreign_column_unique": false
},
{
"table": "account",
"name": "account_id_fkey",
"column": "_type",
"nullable": false,
"unique": false,
"foreign_table": "entity",
"foreign_column": "id",
"foreign_column_nullable": false,
"foreign_column_unique": true
},
{
"table": "account",
"name": "account_id_fkey",
"column": "_type",
"nullable": false,
"unique": false,
"foreign_table": "entity",
"foreign_column": "type",
"foreign_column_nullable": false,
"foreign_column_unique": false
}
],The issue lies within this chunk. Your schema that you've pasted has a composite foreign key and that's something I didn't even know (or forgot) existed and therefore we've never tested for.
In the debug dump you've provided the code has decided instead of a single foreign key, you have 4. And they're all named the same thing which is why we're seeing issues.
In this particular case I'll ask why you've designed your database like this. The entity primary key is id, and the row is uniquely identified by that ID so the accounts table should not have the type on it at all because this is duplication of data and is not a fully normalized design. This is wasteful storage wise but also slower to do joins to the entity table on. Though maybe it's intentional for a reason I'm not comprehending. Can you explain why it's like this?
Another issue may be the uniqueness of the id/type on entity. I think it might mess things up in sqlboiler the way its defined and its defined in an inefficient way anyway. It should simply be a primary key on id and a unique index on type. No composite uniqueness required. The reason is is unique creates an index just like a primary key which takes up space in the db and id is already covered for uniqueness because of the primary key. So inserting a row that would trip that unique index like: 1, "a" and 1, "a" isn't possible already because the id must be unique due to primary key already anyway. So I'd advise you to change that just in general. Unless there's something I'm missing here, I'd be curious what the reasoning is.
Summary is: You have two elements that while possible in postgres aren't currently supported in sqlboiler, I don't understand the use case for the things you've done in your database and so I'd like to work on establishing why these things are useful before I attempt any bug fixing to support this.
dvic
I have an example that produces structs with duplicate fields, e.g.,
What version of SQLBoiler are you using (
sqlboiler --version)?SQLBoiler v3.0.0-rc9If this happened at generation time what was the full SQLBoiler command you used to generate your models? (if not applicable leave blank)
//go:generate sqlboiler --wipe psqlIf this happened at runtime what code produced the issue? (if not applicable leave blank)
What is the output of the command above with the
-dflag added to it? (Provided you are comfortable sharing this, it contains a blueprint of your schema)Please provide a relevant database schema so we can replicate your issue (Provided you are comfortable sharing this)
Further information. What did you do, what did you expect?
The code does not compile.