#sqlx

sqlx is a library which provides a set of extensions on go's standard database/sql library. The sqlx versions of sql.DB, sql.TX, sql.Stmt, et al. all leave the underlying interfaces untouched, so that their interfaces are a superset on the standard ones. This makes it relatively painless to integrate existing codebases using database/sql with sqlx.

Major additional concepts are:

• Marshal rows into structs (with embedded struct support), maps, and slices
• Named parameter support including prepared statements
• Get and Select to go quickly from query to struct/slice
• Common error handling mnemonics (eg. Execf, Execp (MustExec), and Execl)
• LoadFile for executing statements from a file

There is now some fairly comprehensive documentation for sqlx. You can also read the usage below for a quick sample on how sqlx works, or check out the API documentation on godoc.

Important API Stability Note

The sqlx API has been stable for a long time as I have attempted to learn from the way that it interacts with real code and taken in bug reports. There have been very minor breaking changes in the past, many of which would not have affected most code.

However, the API is very large, and there's little evidence that some of the repeated "convenience" verbs deserve to be so widely distributed amongst the nouns of DB, Tx, and Stmt. Removing these methods but preserving the base convenience functions (which take any of these via an interface) would reduce the surface of the API a lot and make it easier to understand.

In addition to this, there are some functions which were exported with the intent to facilitate building richer ORM libraries on top of sqlx, but after having built one myself, I'm not convinced that this is possible given the limited exposure of the name mapping that is revealed in the public API. If it's a choice between exposing more of it or exposing none of it, I'd rather expose none of it in the sqlx namespace.

install

go get github.com/jmoiron/sqlx

issues

Row headers can be ambiguous (SELECT 1 AS a, 2 AS a), and the result of Columns() can have duplicate names on queries like:

SELECT a.id, a.name, b.id, b.name FROM foos AS a JOIN foos AS b ON a.parent = b.id;

making a struct or map destination ambiguous. Use AS in your queries to give rows distinct names, rows.Scan to scan them manually, or SliceScan to get a slice of results.

usage

Below is an example which shows some common use cases for sqlx. Check sqlx_test.go for more usage.

package main

import (
    _ "github.com/lib/pq"
    "database/sql"
    "github.com/jmoiron/sqlx"
    "log"
)

var schema = `
CREATE TABLE person (
    first_name text,
    last_name text,
    email text
);

CREATE TABLE place (
    country text,
    city text NULL,
    telcode integer
)`

type Person struct {
    FirstName string `db:"first_name"`
    LastName  string `db:"last_name"`
    Email     string
}

type Place struct {
    Country string
    City    sql.NullString
    TelCode int
}

func main() {
    // this connects & tries a simple 'SELECT 1', panics on error
    // use sqlx.Open() for sql.Open() semantics
    db, err := sqlx.Connect("postgres", "user=foo dbname=bar sslmode=disable")
    if err != nil {
        log.Fatalln(err)
    }

    // exec the schema or fail; multi-statement Exec behavior varies between
    // database drivers;  pq will exec them all, sqlite3 won't, ymmv
    db.Execf(schema)
    
    tx := db.MustBegin()
    tx.Execl("INSERT INTO person (first_name, last_name, email) VALUES ($1, $2, $3)", "Jason", "Moiron", "[email protected]")
    tx.Execl("INSERT INTO person (first_name, last_name, email) VALUES ($1, $2, $3)", "John", "Doe", "[email protected]")
    tx.Execl("INSERT INTO place (country, city, telcode) VALUES ($1, $2, $3)", "United States", "New York", "1")
    tx.Execl("INSERT INTO place (country, telcode) VALUES ($1, $2)", "Hong Kong", "852")
    tx.Execl("INSERT INTO place (country, telcode) VALUES ($1, $2)", "Singapore", "65")
    // Named queries can use structs, so if you have an existing struct (i.e. person := &Person{}) that you have populated, you can pass it in as &person
    tx.NamedExec("INSERT INTO person (first_name, last_name, email) VALUES (:first_name, :last_name, :email)", &Person{"Jane", "Citizen", "[email protected]"})
    tx.Commit()

    // Query the database, storing results in a []Person (wrapped in []interface{})
    people := []Person{}
    db.Select(&people, "SELECT * FROM person ORDER BY first_name ASC")
    jason, john := people[0], people[1]

    fmt.Printf("%#v\n%#v", jason, john)
    // Person{FirstName:"Jason", LastName:"Moiron", Email:"[email protected]"}
    // Person{FirstName:"John", LastName:"Doe", Email:"[email protected]"}

    // You can also get a single result, a la QueryRow
    jason = Person{}
    err = db.Get(&jason, "SELECT * FROM person WHERE first_name=$1", "Jason")
    fmt.Printf("%#v\n", jason)
    // Person{FirstName:"Jason", LastName:"Moiron", Email:"[email protected]"}

    // if you have null fields and use SELECT *, you must use sql.Null* in your struct
    places := []Place{}
    err := db.Select(&places, "SELECT * FROM place ORDER BY telcode ASC")
    if err != nil {
        fmt.Printf(err)
        return
    }
    usa, singsing, honkers = places[0], places[1], places[2]
    
    fmt.Printf("%#v\n%#v\n%#v\n", usa, singsing, honkers)
    // Place{Country:"United States", City:sql.NullString{String:"New York", Valid:true}, TelCode:1}
    // Place{Country:"Singapore", City:sql.NullString{String:"", Valid:false}, TelCode:65}
    // Place{Country:"Hong Kong", City:sql.NullString{String:"", Valid:false}, TelCode:852}

    // Loop through rows using only one struct
    place := Place{}
    rows, err := db.Queryx("SELECT * FROM place")
    for rows.Next() {
        err := rows.StructScan(&place)
        if err != nil {
            log.Fataln(err)
        } 
        fmt.Printf("%#v\n", place)
    }
    // Place{Country:"United States", City:sql.NullString{String:"New York", Valid:true}, TelCode:1}
    // Place{Country:"Hong Kong", City:sql.NullString{String:"", Valid:false}, TelCode:852}
    // Place{Country:"Singapore", City:sql.NullString{String:"", Valid:false}, TelCode:65}

    // Named queries, using `:name` as the bindvar.  Automatic bindvar support
    // which takes into account the dbtype based on the driverName on sqlx.Open/Connect
    _, err = db.NamedExec(`INSERT INTO person (first_name,last_name,email) VALUES (:first,:last,:email)`, 
        map[string]interface{}{
            "first": "Bin",
            "last": "Smuth",
            "email": "[email protected]",
    })

    // Selects Mr. Smith from the database
    rows, err := db.NamedQuery(`SELECT * FROM person WHERE first_name=:fn`, map[string]interface{}{"fn": "Bin"})

    // Named queries can also use structs.  Their bind names follow the same rules
    // as the name -> db mapping, so struct fields are lowercased and the `db` tag
    // is taken into consideration.
    rows, err := db.NamedQuery(`SELECT * FROM person WHERE first_name=:first_name`, jason)
}

embedded structs

Structs which do not implement the sql.Scanner interface will be inspected and their fields used as possible targets for a scan. This includes embedded and non-embedded structs.

Go makes 'ambiguous selectors' a compile time error, but does not make structs with possible ambiguous selectors errors. Sqlx will decide which field to use on a struct based on a breadth first search of the struct and any structs it contains or embeds, as specified by the order of the fields as accessible by reflect, which generally means in source-order.

scan safety

By default, scanning into structs requires the structs to have fields for all of the columns in the query. This was done for a few reasons:

• A mistake in naming during development could lead you to believe that data is being written to a field when actually it can't be found and it is being dropped
• This behavior mirrors the behavior of the Go compiler with respect to unused variables
• Selecting more data than you need is wasteful (more data on the wire, more time marshalling, etc)

Unlike Marshallers in the stdlib, the programmer scanning an sql result into a struct will generally have a full understanding of what the underlying data model is and full control over the SQL statement.

Despite this, there are use cases where it's convenient to be able to ignore unknown columns. In most of these cases, you might be better off with ScanSlice, but where you want to still use structs, there is now the Unsafe method. Its usage is most simply shown in an example:

    db, err := sqlx.Connect("postgres", "user=foo dbname=bar sslmode=disable")
    if err != nil {
        log.Fatal(err)
    }

    type Person {
        Name string
    }
    var p Person

    // This fails, because there is no destination for location in Person
    err = db.Get(&p, "SELECT name, location FROM person LIMIT 1")
    
    udb := db.Unsafe()
    
    // This succeeds and just sets `Name` in the p struct
    err = udb.Get(&p, "SELECT name, location FROM person LIMIT 1")

The Unsafe method is implemented on Tx, DB, and Stmt. When you use an unsafe Tx or DB to create a new Tx or Stmt, those inherit its lack of safety.