README
¶
sqan
Dead simple package for scanning SQL rows, compatible with the standard library database/sql.
Unlike other libraries, sqan maps structs recursively, this simplifies the scanning of complex queries and allows to handle data in the same table with different (child) objects.
Install
go get -u github.com/GGP1/sqan
Usage
type User struct {
ID int `json:"id"`
Name string `json:"name"`
Stats Stats `json:"stats"`
CreatedAt time.Time `json:"created_at" db:"created_at"`
}
type Stats struct {
// Other libraries wouldn't map this field
FriendsCount int `json:"friends_count" db:"friends_count"`
}
// Error handling omitted for simplicity sake
func main() {
rows, _ := db.Query("SELECT * FROM users")
var users []User
_ = sqan.Rows(&users, rows)
for _, user := range users {
fmt.Println(user.ID, user.Name, user.Stats.FriendsCount, user.CreatedAt)
}
// 1 "Alice" 4 time.Time
// 2 "Bob" 3 time.Time
rows, _ = db.Query("SELECT name, friends_count FROM users WHERE id=$1", 1)
var user User
_ = sqan.Row(&user, rows)
fmt.Println(user.ID, user.Name, user.Stats.FriendsCount, user.CreatedAt)
// In some cases, it may be useful to use pointer fields so the empty ones are null
// 0 "Alice" 4 time.Time{}
}
Documentation
Mapping
Unexported fields and struct slices aren't mapped.
The "db" tag can be used to map a struct field with an SQL one, if no tag is used, the mapping is done by converting the field's name to lower case.
Objects are mapped only once and the mapping is kept inside a Go map for later use. It is assumed that the number of objects to map is not high enough to cause memory issues.
The sqan.Row function takes sql.Rows as it's not possible to access the returned columns and map them through sql.Row.
Documentation
Source Files