Skip to content

Indexes

Indexes

dbx supports schema-level index declaration. Migration planning and AutoMigrate create missing indexes based on schema metadata.

When to Use

  • You want schema-owned index definitions.
  • You need both single-column and composite indexes.

Single-Column 

type UserSchema struct {
	dbx.Schema[User]
	Username dbx.Column[User, string] `dbx:"username,index"`
	Email    dbx.Column[User, string] `dbx:"email,unique"`
}

Composite 

type UserSchema struct {
	dbx.Schema[User]
	TenantID dbx.Column[User, int64]  `dbx:"tenant_id"`
	Username dbx.Column[User, string] `dbx:"username"`
	Email    dbx.Column[User, string] `dbx:"email"`

	ByTenantAndUsername dbx.Index[User]  `idx:"columns=tenant_id|username"`
	UniqueTenantEmail   dbx.Unique[User] `idx:"columns=tenant_id|email"`
}

Naming

  • primary key: pk_<table>
  • non-unique index: idx_<table>_<columns>
  • unique index: ux_<table>_<columns>

Pitfalls

  • idx:"columns=..." must use column names from schema tags.
  • Too many write-heavy indexes can reduce insert/update throughput.

Verify 

go test ./dbx/... -run Migrate