Entity generation for partitioned tables

[jared-mackey]

Description

Hi,

I'm not certain this is a bug but I don't know if it's correct behavior either. When generating entities for my schema I noticed that the entity generator does a few unexpected things with my partitioned tables:

• It tries to generate an entity for each partition instead of just one for the super table.
  • I was able to ignore these though via the CLI parameters.
• Entity relations are being created for table that FK to the super table but one per partition.

Steps to Reproduce

1. Create a partitioned table. Let's call this table_1 and then each partition will be table_1_p1 through table_1_p128.
2. Create a normal table that FKs to the partition table. Let's call this table_2.
3. Generate the entities with sea-orm-cli generate entity -o entity/src --lib --tables "table_1,table_2" --ignore-tables "table_1_p1,table_1_p2..." --max-connections 25

Expected Behavior

• I would not get entities created for table_1_p1..table_1_p128.
• Entity for table_2 would have a single relationship defined -- the one to table_1.

Actual Behavior

✅ I did not get entities created for table_1_p1
✅ The table_2 entity correctly had a FK to the main table.
❌ The table_2 entity was filled with relationships to table_1_p1..table_1_p128.

Example

#[derive(Copy, Clone, Debug, EnumIter, DeriveRelation)]
pub enum Relation {
    #[sea_orm(
        belongs_to = "super::table_1_p34::Entity", // <-- actually doesn't exist because I don't generate those entities
        from = "Column::Table1Id",
        to = "super::table_1_p34::Column::Id", // <-- same
        on_update = "NoAction",
        on_delete = "NoAction"
    )]
    Table1P34,
...
}

impl Related<super::table_2_p34::Entity> for Entity {
    fn to() -> RelationDef {
        Relation::Table1P34.def()
    }
}

Reproduces How Often

Every time.

Workarounds

Delete the wrongly generated code.

Versions

0.11.2

[tyt2y3]

Can you create a new folder under /issues, attach a minimal schema and open a PR?
Schema discovery do not take in consideration partitioned tables, so we'll need to figure out how to identify and remove the duplicates.

[cemremengu]

As a note, other libraries like jooq (java) provide a regex mechanism to exclude unwanted tables like partitions in a schema such as

<excludes>.*_p[0-9_]+</excludes>

[billy1624]

Hey @jared-mackey, thanks for the report! I just drafted a PR to skip parsing partitioned Postgres tables.

[billy1624]

The patch will be released in the coming 0.12.x

[github-actions]

🎉 Released In 0.12.1 🎉

Thank you everyone for the contribution!
This feature is now available in the latest release. Now is a good time to upgrade!
Your participation is what makes us unique; your adoption is what drives us forward.
You can support SeaQL 🌊 by starring our repos, sharing our libraries and becoming a sponsor ⭐.