01-basic-entity/

01-basic-entity

One User entity, written out in every field type and every query type. Start here.

• Every field type: String, Int, Int64, Float, Bool, Byte, Time, JSON
• Field options: Unique(), Optional(), Immutable(), Default(), DefaultFunc()
• Validate() with your own Go function from the logic package
• Field level Contracts(): a password that clients write but never read, timestamps they read but never write
• Ready made queries: DefaultCRUD(), CreateBulk(), ListAll(), DeleteAll()
• CreateBulk().Upsert("email"): re-importing a row overwrites it instead of failing on the unique email
• Queries by field: GetBy("email"), ListBy("is_active"), and Name() to rename one
• Filters: filter.Range() and filter.Search()
• Indexes: multi column index.Asc() / index.Desc(), mixed sort order, Unique(), Name()
• The same schema on three dialects: sqlite, postgresql, mysql

Entity

ent/schema/user.go is the same file in all three dialect folders. Only sqlc.yaml and the generated SQL differ.

Fields() lists one field per type, so you can see what each one becomes in schema.sql and in schema.proto. Two fields are worth a look:

• password is SQLC() plus PROTO().WriteOnly(). It is a column, and clients can send it, but it is not in any response message.
• created_at and updated_at are PROTO().ReadOnly(). The server sets them with DefaultFunc(time.Now), so they are not in create or update requests.

api_key shows DefaultFunc on bytes. The value is generated in Go, not by the database.

Dialects

Compare the three ent/contract/sqlc/schema.sql files to see how types map. For example field.Float becomes REAL on sqlite, DOUBLE PRECISION on postgresql and DOUBLE on mysql.

Run

cd sqlite
make run     # generates types, bundles the frontend, serves on :8080

For postgresql and mysql, make down stops the database and drops its data.