The db.Record interface

The db.Record interface provides a bare minimum support for you to operate with Go structs in a more ORM-ish way:

package db

type Record interface {
  Store(sess Session) Store
}

The db.Store interface is a container for db.Collection that is defined like this:

type Store interface {
  Collection
}

In the example below, Book is a struct that satisfies the db.Record interface:

package main

import (
  "github.com/upper/db/v4"
)

type Book struct {
  ID        uint   `db:"id,omitempty"`
  Title     string `db:"title"`
  AuthorID  uint   `db:"author_id,omitempty"`
  SubjectID uint   `db:"subject_id,omitempty"`
}

func (book *Book) Store(sess db.Session) db.Store {
  return sess.Collection("books")
}

// Compile-time check.
var _ = db.Record(&Book{})

You can use the db.Record interface with special db.Session methods such as Get, Save or Delete:

var book Book

// Get "The Shining" from the catalog
err = sess.Get(&book, db.Cond{"title": "The Shining"})
// ...

// Persist record to database
err = sess.Save(&book)
// ...

// Create a new record
book := Book{Title: "My new book"}
err = sess.Save(&book)

// Delete record
err = sess.Delete(&book)

Hooks

db.Record objects can optionally satisfy hooks, which are special methods called before or after specific events. For instance, if we’d like the Book record to execute code right before inserting a new entry into the database, we’d add a BeforeCreate hook like this:

func (book *Book) BeforeCreate(sess db.Session) error {
  // ...
  return nil
}

upper/db records support the following hooks:

• BeforeCreate(db.Session) error
• AfterCreate(db.Session) error
• BeforeUpdate(db.Session) error
• AfterUpdate(db.Session) error
• BeforeDelete(db.Session) error
• AfterDelete(db.Session) error

Hooks in upper/db run within a database transaction; if any of the hooks returns an error, the whole operation is canceled and rolled back.

Validation

Besides hooks, there’s another optional interface defined as:

type Validator interface {
  Validate() error
}

You can use the Validator interface to run validations against the record’s data.