dat

There are ORMs available for Golang, but my experience has been better with packages like sqlx [1] or dat [2]. We've since been using SQLX for pretty much all DBMS related work, and regret nothing. In my experience SQLX gave us the right balance between abstraction and control. [1] http://jmoiron.github.io/sqlx/ [2] https://github.com/mgutz/dat

sqlx (https://github.com/jmoiron/sqlx) is my go-to as well. mgutz/dat (https://github.com/mgutz/dat) and dbr (https://github.com/gocraft/dbr) are also more ORM-like alternatives if that's what you need—query builders with the option to use raw SQL for more complex queries. I also lean on purse (https://github.com/smotes/purse) for keeping my SQL in separate files, but I'd love to see something like Clojure's yesql. I think the likelihood of seeing an ActiveRecord-style ORM in Go is slim: ActiveRecord is a core part of Rails and can therefore make a lot of safe assumptions about its environment. Any ORM for Go can't do that to the same extent, and big Rails-style frameworks don't gain much of a following in Go.

Postgres 9.3+ is fairly straight-forward too. Here is go + github.com/mgutz/dat // one trip to database using subqueries and Postgres' JSON functions con.SelectDoc("id", "user_name", "avatar"). HasMany("recent_comments", `SELECT id, title FROM comments WHERE id = users.id LIMIT 10`). HasMany("recent_posts", `SELECT id, title FROM posts WHERE author_id = users.id LIMIT 10`). HasOne("account", `SELECT balance FROM accounts WHERE user_id = users.id`). From("users"). Where("id = $1", 4). QueryStruct(&obj) // obj must be agreeable with json.Unmarshal() results in { "id": 4, "user_name": "mario", "avatar": "https://imgur.com/a23x.jpg", "recent_comments": [{"id": 1, "title": "..."}], "recent_posts": [{"id": 1, "title": "..."}], "account":