Support for custom implementations of Scan for models

[azaviyalov]

I'm moving from sqlx to scany because of using pgx as a postgres driver.

I have a custom implementation of sql.Scanner for my model, but it seems doesn't working with scany.

func (m *MyModel) Scan(data interface{}) error {
    // ...
}

When selecting rows, I got errors like that: scany: column: 'body': no corresponding field found, or it's unexported in model.MyModel, because MyModel does not really have a body column, but a custom Scan instead.

Is it possible to implement check if models implement some Scanner interface?

[georgysavva]

Hi. Yes, it should work with scany. Could you please provide the code where you define your struct, how you pass it to scan and the SQL query?

[azaviyalov]

Yes, sure

package main

import (
    "fmt"
    "context"

    "github.com/jackc/pgx/v4/pgxpool"
    "github.com/georgysavva/scany/pgxscan"
)

type User struct {
    // no columns defined here
}

/*
create table users
(
    id serial constraint users_pk primary key,
    name varchar not null
);

create unique index users_id_uindex on users (id);

insert into users (id, name)
values  (1, 'john'),
        (2, 'jacob'),
        (3, 'joshua');

*/

func (m *User) Scan(data interface{}) error {
    // seems not to be called
    // scany should not have any errors, if it is
    return nil
}

func main() {
    ctx := context.Background()
    db, _ := pgxpool.Connect(ctx, "postgres://postgres:mysecretpassword@localhost:5432/postgres")

    var users []*User
    err := pgxscan.Select(ctx, db, &users, `SELECT id, name FROM users`)
    if err != nil {
        panic(err)
    }

    fmt.Println(users)
}

output is panic: scany: column: 'id': no corresponding field found, or it's unexported in main.User

[georgysavva]

Scany supports custom types that implement sql.Scanner interface (see this for more details on how sql.Scanner works with database libraries). Here is the corresponding scany docs section: https://pkg.go.dev/github.com/georgysavva/scany@v0.2.9/dbscan#hdr-NULLs_and_custom_types. Generally, they should be used as fields, not as a top-level model, because scany just passes them to the underlying database library for the corresponding field and the database library handles the custom .Scan() call. For example:

type User struct {
    ID string `db:"id"`
    Name MyString `db:"name"`
}

type MyString struct {
    text string
}

func (ms *MyString) Scan(data interface{}) error {
    // Fill ms.text and do other stuff here.
    return nil
}

func main() {
    ctx := context.Background()
    db, _ := pgxpool.Connect(ctx, "postgres://postgres:mysecretpassword@localhost:5432/postgres")

    var users []*User
    err := pgxscan.Select(ctx, db, &users, `SELECT id, name FROM users`)
    if err != nil {
        panic(err)
    }

    fmt.Println(users)
}

However it could be used as you intended but only for the single-column case, because calling the custom .Scan() method only makes sense on columns:

type MyString struct {
    text string
}

func (ms *MyString) Scan(data interface{}) error {
    // Fill ms.text and do other stuff here.
    return nil
}

func main() {
    ctx := context.Background()
    db, _ := pgxpool.Connect(ctx, "postgres://postgres:mysecretpassword@localhost:5432/postgres")

    var strings []MyString
    err := pgxscan.Select(ctx, db, &strings, `SELECT name FROM users`)
    if err != nil {
        panic(err)
    }

    fmt.Println(strings)
}

Note that the later example won't work with the latest scany version. It's in the master branch and hasn't been released yet.

[azaviyalov]

Thank you, the problem is solved!