gosexy/db is a set of wrappers for SQL and NoSQL database drivers for the Go programming languaje. Currently compatible with MongoDB, MySQL, PostgreSQL and SQLite3.

The goal of this package is to provide a common, simple, consistent layer of abstraction for performing mundane operations such as create, read, update, and delete rows (CRUD) on different databases.

While gosexy/db is not an ORM per se it can be used as the base for one, we leave that up to you, gosexy/db prefers to stay out of the way and just focusing in providing compatibility between databases, it's like a babelfish!

Let me show you an example, this chunk of code searches on the "people" table/collection. It does not matter whether we are querying a NoSQL database like MongoDB or a SQL database like MySQL, PostgreSQL or SQLite3, gosexy/db talks to the database in the language the database expects and returns you a set of results.

items, err := people.FindAll(
  db.Cond{"name": "john"},
)

for i, item := range items {
  // ...
}

db.Collection.FindAll() will accept different structures in no special order, in the above example we are passing a db.Cond{} type, that's a condition, you could also use db.And{}, db.Or{}, db.Limit(n), db.Offset(n), etc. each one of them having different meanings.

While this level of abstraction would not be able to represent a complex query or to use any database-specific features it's fairly convenient for doing the simple CRUD stuff, and regarding more advanced queries, the underlying driver could always be retrieved as a *sql.DB or a *mgo.Session so you can still be able to use any database-pro spells.

Fetching all rows may be not so adequate for processing large datasets, in that case we can use db.Collection.Query() instead of db.Collection.FindAll() and then iterate over the results.

// Makes a query and stores the result.
res, err = people.Query(
  db.Cond{"name": "john"},
)

if err != nil {
  panic(err.Error())
}

person := struct{
  PersonName string `field:"name"` // Supports struct tags.
}{}

for {
  // res.Next() accepts a pointer to map or a pointer to struct.
  err = res.Next(&person)

  if err == nil {
    // If there is no error we can use person.
    fmt.Printf("%v\n", person)
  } else if err == db.ErrNoMoreRows {
    // This error means we have read all rows.
    break
  } else {
    // Another kind of error needs proper management.
    panic(err.Error())
  }

}

One of the features you may find convenient is the ability of gosexy/db to make relations between different databases that talk different protocols with ease:

items, err := peopleCollection.FindAll(
  db.RelateAll{
    "works": db.On{
      worksCollection,
      db.Cond{"author_id": "{id}"},
    },
  },
)

In the above example, peopleCollection and worksCollection are db.Collection objects and they could be collections or tables of any of the supported databases. You can even relate NoSQL collections to SQL tables!

gosexy/db is a work in progress but its core features are ready for use.

Use go get to download and install gosexy/db.

go get menteslibres.net/gosexy/db

The gosexy/db package provides interface definitions and datatypes only, it can't connect to any database by itself, in order to connect to an actual database a database wrapper is required.

Database wrappers are all different and may have special installation requirements, please refer to the appropriate documentation reference on the following list.

• mongo
• mysql
• postgresql
• sqlite

Let's suppose we want to use the "mongo" wrapper for MongoDB.

Use go get to retrieve and install the "mongo" wrapper.

go get menteslibres.net/gosexy/db/mongo

Import gosexy/db and the wrapper into your project.

import (
  "menteslibres.net/gosexy/db"
  // The wrapper goes to the blank identifier.
  _ "menteslibres.net/gosexy/db/mongo"
)

Set up a variable to configure your database connection.

var settings = db.DataSource{
  Host:     "localhost",
  Database: "dbname",
  User:     "myusername",
  Password: "mysecret",
}

Use db.Open to connect to the database you've just set up.

// Connect using the mongo driver.
sess, err := db.Open("mongo", settings)

if err != nil {
  panic(err)
}

defer sess.Close()

Insert some items and retrieve rows.

animals, _ := sess.Collection("animals")

animals.Append(db.Item{
  "animal": "Bird",
  "young":  "Chick",
  "female": "Hen",
  "male":   "Cock",
  "group":  "flock",
})

animals.Append(db.Item{
  "animal": "Bovidae",
  "young":  "Calf",
  "female": "Cow",
  "male":   "Bull",
  "group":  "Herd",
})

animals.Append(db.Item{
  "animal": "Canidae",
  "young":  []string{"Puppy", "Pup"},
  "female": "Bitch",
  "male":   "Dog",
  "group":  "Pack",
})

items, err := animals.FindAll()

if err != nil {
  panic(err.Error())
}

for _, item := range items {
  fmt.Printf("animal: %s, young: %s\n", item["animal"], item["young"])
}

The same example goes for other wrappers, you just change the driver name to mysql, postgresql or sqlite.

We have an online reference and you can always have a quick look to the API at our documentation page.

Got problems? try to use the forum.

This is an evolving project, there are still lots of things to do:

• Add db.Upsert and db.Modify for SQL databases.
• Add CouchDB support.

2013/04/15 - Adding support for Update() with structs and maps.
2013/04/06 - Adding support for struct tags.
2013/03/10 - Breaking change: Find() and FindAll() now will return ([item],
             error).
           - Adding db.Result and methods for iterating over results.
           - Adding the ability to use struct{} for fetching and inserting
             rows.
           - Adding helper package gosexy/db/util for internal usage.
2012/12/02 - Changing db.Table.Collection and adding db.Open().
2012/09/21 - Changing some methods parameters and return values, improving
             error handling and testing many data types.
2012/08/29 - Created the main site docs and moved the repo to
             "http://menteslibres.net/gosexy".
2012/07/23 - Splitted database wrappers into packages. Changed db.Where to
             db.Cond.
2012/07/09 - First public beta with MySQL, MongoDB, PostgreSQL and SQLite3.