delete_by

delete_by_{filter} — same filter grammar as first_by:

async fn delete_by_id(&self, id: i32) -> sqlx::Result<()>;
// DELETE FROM "user" WHERE id = $1

Always returns sqlx::Result<()> — the row count isn't surfaced.

Cannot filter through a relation. Unlike first_by_*, a delete_by_* filter can't reference a joined table's columns at all (even belongs_to's non-JOIN FK shortcut is fine, but anything requiring an actual JOIN is rejected) — SQL's DELETE doesn't support joining against other tables the way SELECT/UPDATE can here. This is a compile error:

#[repository(table = "user")]
#[belongs_to(table = "team")]
pub trait UserRepository: Repository {
    // ERROR: `delete_by_*` cannot filter through joined tables
    async fn delete_by_team_name(&self, name: String) -> sqlx::Result<()>;
    // fine — team_id is the belongs_to FK column itself, no JOIN needed
    async fn delete_by_team_id(&self, team_id: i32) -> sqlx::Result<()>;
}

For a delete that genuinely needs to filter on a joined column, write the method body by hand (e.g. DELETE FROM "user" WHERE id IN (SELECT ...)).

Not supported here: #[order_by(...)] (a compile error if present).