License GoDoc CircleCI Go Report Card

SQLBoiler is a tool to generate a Go ORM tailored to your database schema.

It is a "database-first" ORM as opposed to "code-first" (like gorm/gorp). That means you must first create your database schema. Please use something like goose, sql-migrate or some other migration tool to manage this part of the database's life-cycle.

Why another ORM

While attempting to migrate a legacy Rails database, we realized how much ActiveRecord benefitted us in terms of development velocity. Coming over to the Go database/sql package after using ActiveRecord feels extremely repetitive, super long-winded and down-right boring. Being Go veterans we knew the state of ORMs was shaky, and after a quick review we found what our fears confirmed. Most packages out there are code-first, reflect-based and have a very weak story around relationships between models. So with that we set out with these goals:

  • Work with existing databases: Don't be the tool to define the schema, that's better left to other tools.
  • ActiveRecord-like productivity: Eliminate all sql boilerplate, have relationships as a first-class concept.
  • Go-like feel: Work with normal structs, call functions, no hyper-magical struct tags, small interfaces.
  • Go-like performance: Benchmark and optimize the hot-paths, perform like hand-rolled sql.DB code.

We believe with SQLBoiler and our database-first code-generation approach we've been able to successfully meet all of these goals. On top of that SQLBoiler also confers the following benefits:

  • The models package is type safe. This means no chance of random panics due to passing in the wrong type. No need for interface{}.
  • Our types closely correlate to your database column types. This is expanded by our extended null package which supports nearly all Go data types.
  • A system that is easy to debug. Your ORM is tailored to your schema, the code paths should be easy to trace since it's not all buried in reflect.
  • Auto-completion provides work-flow efficiency gains.

Table of Contents

About SQL Boiler

  • Full model generation
  • High performance through generation
  • Extremely fast code generation
  • Uses boil.Executor (simple interface, sql.DB, sqlx.DB etc. compatible)
  • Easy workflow (models can always be regenerated, full auto-complete)
  • Strongly typed querying (usually no converting or binding to pointers)
  • Hooks (Before/After Create/Select/Update/Delete/Upsert)
  • Automatic CreatedAt/UpdatedAt
  • Relationships/Associations
  • Eager loading (recursive)
  • Transactions
  • Raw SQL fallback
  • Compatibility tests (Run against your own DB schema)
  • Debug logging
Supported Databases
  • PostgreSQL

Note: Seeking contributors for other database engines.

A Small Taste

For a comprehensive list of available operations and examples please see Features & Examples.

import (
  // Import this so we don't have to use qm.Limit etc.
  . ""

// Open handle to database like normal
db, err := sql.Open("postgres", "dbname=fun user=abc")
if err != nil {
  return err

// Query all users
users, err := models.Users(db).All()

// Panic-able if you like to code that way
users := models.Users(db).AllP()

// More complex query
users, err := models.Users(db, Where("age > ?", 30), Limit(5), Offset(6)).All()

// Ultra complex query
users, err := models.Users(db,
  Select("id", "name"),
  InnerJoin("credit_cards c on c.user_id ="),
  Where("age > ?", 30),
  AndIn("c.kind in ?", "visa", "mastercard"),
  Or("email like ?", ""),
  GroupBy("id", "name"),
  Having("count( > ?", 2),

// Use any "boil.Executor" implementation (*sql.DB, *sql.Tx, data-dog mock db)
// for any query.
tx, err := db.Begin()
if err != nil {
  return err
users, err := models.Users(tx).All()

// Relationships
user, err := models.Users(db).One()
if err != nil {
  return err
movies, err := user.FavoriteMovies(db).All()

// Eager loading
users, err := models.Users(db, Load("FavoriteMovies")).All()
if err != nil {
  return err

Requirements & Pro Tips

  • Go 1.6 minimum, and Go 1.7 for compatibility tests.
  • Table names and column names should use snake_case format.
    • At the moment we require snake_case table names and column names. This is a recommended default in Postgres. We can reassess this for future database drivers.
  • Join tables should use a composite primary key.
    • For join tables to be used transparently for relationships your join table must have a composite primary key that encompasses both foreign table foreign keys. For example, on a join table named user_videos you should have: primary key(user_id, video_id), with both user_id and video_id being foreign key columns to the users and videos tables respectively.
Pro Tips
  • Foreign key column names should end with _id.
    • Foreign key column names in the format x_id will generate clearer method names. It is advisable to use this naming convention whenever it makes sense for your database schema.
  • If you never plan on using the hooks functionality you can disable generation of this feature using the --no-hooks flag. This will save you some binary size.

Getting started

go get -u -t

Create a configuration file. Because the project uses viper, TOML, JSON and YAML are all supported. Environment variables are also able to be used. We will assume TOML for the rest of the documentation.

The configuration file should be named sqlboiler.toml and is searched for in the following directories in this order:

  • ./
  • $XDG_CONFIG_HOME/sqlboiler/
  • $HOME/.config/sqlboiler/

We require you pass in the postgres configuration via the configuration file rather than env vars. There is no command line argument support for database configuration. Values given under the postgres block are passed directly to the pq driver. Here is a rundown of all the different values that can go in that section:

Name Required Default
dbname yes none
host yes none
port no 5432
user yes none
pass no none
sslmode no "require"

You can also pass in these top level configuration values if you would prefer not to pass them through the command line or environment variables:

Name Default
basedir none
pkgname "models"
output "models"
exclude [ ]
tags [ ]
debug false
no-hooks false
no-tests false
no-auto-timestamps false


Initial Generation

After creating a configuration file that points at the database we want to generate models for, we can invoke the sqlboiler command line utility.

SQL Boiler generates a Go ORM from template files, tailored to your database schema.
Complete documentation is available at

  sqlboiler [flags] <driver>

sqlboiler postgres

  -b, --basedir string        The base directory has the templates and templates_test folders
  -d, --debug                 Debug mode prints stack traces on error
  -x, --exclude stringSlice   Tables to be excluded from the generated package
      --no-auto-timestamps    Disable automatic timestamps for created_at/updated_at
      --no-hooks              Disable hooks feature for your models
      --no-tests              Disable generated go test files
  -o, --output string         The name of the folder to output to (default "models")
  -p, --pkgname string        The name you wish to assign to your generated package (default "models")
  -t, --tag stringSlice       Struct tags to be included on your models in addition to json, yaml, toml

Follow the steps below to do some basic model generation. Once we've generated our models, we can run the compatibility tests which will exercise the entirety of the generated code. This way we can ensure that our database is compatible with SQLBoiler. If you find there are some failing tests, please check the Diagnosing Problems section.

# Generate our models and exclude the migrations table
sqlboiler -x goose_migrations postgres

# Run the generated tests
go test ./models # This requires an administrator postgres user because of some
                 # voodoo we do to disable triggers for the generated test db

Diagnosing Problems

The most common causes of problems and panics are:

  • Forgetting to exclude tables you do not want included in your generation, like migration tables.
  • Tables without a primary key. All tables require one.
  • Forgetting to put foreign key constraints on your columns that reference other tables.
  • The compatibility tests that run against your own DB schema require a superuser, ensure the user supplied in your sqlboiler.toml config has adequate privileges.
  • A nil or closed database handle. Ensure your passed in boil.Executor is not nil.
    • If you decide to use the G variant of functions instead, make sure you've initialized your global database handle using boil.SetDB().

For errors with other causes, it may be simple to debug yourself by looking at the generated code. Setting boil.DebugMode to true can help with this. You can change the output using boil.DebugWriter (defaults to os.Stdout).

If you're still stuck and/or you think you've found a bug, feel free to leave an issue and we'll do our best to help you.

Features & Examples

Most examples in this section will be demonstrated using the following schema, structs and variables:

  id integer NOT NULL,
  name text NOT NULL,


  id integer NOT NULL,
  pilot_id integer NOT NULL,
  age integer NOT NULL,
  name text NOT NULL,
  color text NOT NULL,

ALTER TABLE jets ADD CONSTRAINT pilots_fkey FOREIGN KEY (pilot_id) REFERENCES pilots(id);

CREATE TABLE languages (
  id integer NOT NULL,
  language text NOT NULL

ALTER TABLE languages ADD CONSTRAINT language_pkey PRIMARY KEY (id);

-- Join table
CREATE TABLE pilot_languages (
  pilot_id integer NOT NULL,
  language_id integer NOT NULL

-- Composite primary key
ALTER TABLE pilot_languages ADD CONSTRAINT pilot_language_pkey PRIMARY KEY (pilot_id, language_id);
ALTER TABLE pilot_languages ADD CONSTRAINT pilots_fkey FOREIGN KEY (pilot_id) REFERENCES pilots(id);
ALTER TABLE pilot_languages ADD CONSTRAINT languages_fkey FOREIGN KEY (language_id) REFERENCES languages(id);

The generated model structs for this schema look like the following. Note that I've included the relationship structs as well so you can see how it all pieces together, but these are unexported and not something you should ever need to touch directly:

type Pilot struct {
  ID   int    `boil:"id" json:"id" toml:"id" yaml:"id"`
  Name string `boil:"name" json:"name" toml:"name" yaml:"name"`

  R *pilotR `boil:"-" json:"-" toml:"-" yaml:"-"`

type pilotR struct {
  Licenses  LicenseSlice
  Languages LanguageSlice
  Jets      JetSlice

type Jet struct {
  ID      int    `boil:"id" json:"id" toml:"id" yaml:"id"`
  PilotID int    `boil:"pilot_id" json:"pilot_id" toml:"pilot_id" yaml:"pilot_id"`
  Age     int    `boil:"age" json:"age" toml:"age" yaml:"age"`
  Name    string `boil:"name" json:"name" toml:"name" yaml:"name"`
  Color   string `boil:"color" json:"color" toml:"color" yaml:"color"`

  R *jetR `boil:"-" json:"-" toml:"-" yaml:"-"`

type jetR struct {
  Pilot *Pilot

type Language struct {
  ID       int    `boil:"id" json:"id" toml:"id" yaml:"id"`
  Language string `boil:"language" json:"language" toml:"language" yaml:"language"`

  R *languageR `boil:"-" json:"-" toml:"-" yaml:"-"`

type languageR struct {
  Pilots PilotSlice
// Open handle to database like normal
db, err := sql.Open("postgres", "dbname=fun user=abc")
if err != nil {
  return err
Automatic CreatedAt/UpdatedAt

If your generated SQLBoiler models package can find columns with the names created_at or updated_at it will automatically set them to time.Now() in your database, and update your object appropriately. To disable this feature use --no-auto-timestamps.

Note: You can set the timezone for this feature by calling boil.SetLocation()

Overriding Automatic Timestamps
  • Insert
    • Timestamps for both updated_at and created_at that are zero values will be set automatically.
    • To set the timestamp to null, set Valid to false and Time to a non-zero value. This is somewhat of a work around until we can devise a better solution in a later version.
  • Update
    • The updated_at column will always be set to time.Now(). If you need to override this value you will need to fall back to another method in the meantime: boil.SQL(), overriding updated_at in all of your objects using a hook, or create your own wrapper.
  • Upsert
    • created_at will be set automatically if it is a zero value, otherwise your supplied value will be used. To set created_at to null, set Valid to false and Time to a non-zero value.
    • The updated_at column will always be set to time.Now().
Query Building

We generate "Starter" methods for you. These methods are named as the plural versions of your model, for example: models.Jets(). Starter methods are used to build queries using our Query Mod System. They take a slice of Query Mods as parameters, and end with a call to a Finisher method.

Here are a few examples:

// SELECT COUNT(*) FROM pilots;
count, err := models.Pilots().Count()

// SELECT * FROM "pilots" LIMIT 5;
pilots, err := models.Pilots(qm.Limit(5)).All()

// DELETE FROM "pilots" WHERE "id"=$1;
err := models.Pilots(qm.Where("id=?", 1)).DeleteAll()

In the event that you would like to build a query and specify the table yourself, you can do so using models.NewQuery():

// Select all rows from the pilots table by using the From query mod.
err := models.NewQuery(db, From("pilots")).All()

As you can see, Query Mods allow you to modify your queries, and Finishers allow you to execute the final action.

If you plan on executing the same query with the same values using the query builder, you should do so like the following to utilize caching:

// Instead of this:
for i := 0; i < 10; i++ {
   pilots := models.Pilots(qm.Where("id > ?", 5), qm.Limit(5)).All()

// You should do this
query := models.Pilots(qm.Where("id > ?", 5), qm.Limit(5))
for i := 0; i < 10; i++ {
   pilots := query.All()

// Every execution of All() after the first will use a cached version of
// the built query that short circuits the query builder all together.
// This allows you to save on performance.

// Just something to be aware of: query mods don't store pointers, so if
// your passed in variable's value changes, your generated query will not change.

Note: You will see exported boil.SetX methods in the boil package. These should not be used on query objects because they will break caching. Unfortunately these had to be exported due to some circular dependency issues, but they're not functionality we want exposed. If you want a different query object, generate a new one.

Take a look at our Relationships Query Building section for some additional query building information.

Query Mod System

The query mod system allows you to modify queries created with Starter methods when performing query building. Here is a list of all of your generated query mods using examples:

// Dot import so we can access query mods directly instead of prefixing with "qm."
import . ""

// Use a raw query against a generated struct (Pilot in this example)
// If this query mod exists in your call, it will override the others.
// "?" placeholders are not supported here, use "$1, $2" etc.
SQL("select * from pilots where id=$1", 10)
models.Pilots(SQL("select * from pilots where id=$1", 10)).All()

Select("id", "name") // Select specific columns.
From("pilots as p") // Specify the FROM table manually, can be useful for doing complex queries.

// WHERE clause building
Where("name=?", "John")
And("age=?", 24)
Or("height=?", 183)

// WHERE IN clause building
WhereIn("name, age in ?", "John" 24, "Tim", 33) // Generates: WHERE ("name","age") IN (($1,$2),($3,$4))
AndIn("weight in ?", 84)
OrIn("height in ?", 183, 177, 204)

InnerJoin("pilots p on jets.pilot_id=?", 10)

OrderBy("age, height")

Having("count(jets) > 2")


// Explicit locking
For("update nowait")

// Eager Loading -- Load takes the relationship name, ie the struct field name of the
// Relationship struct field you want to load.
Load("Languages") // If it's a ToOne relationship it's in singular form, ToMany is plural.

Note: We don't force you to break queries apart like this if you don't want to, the following is also valid and supported by query mods that take a clause:

Where("(name=? OR age=?) AND height=?", "John", 24, 183)
Function Variations

You will find that most functions have the following variations. We've used the Delete method to demonstrate:

// Set the global db handle for G method variants.

pilot, _ := models.FindPilot(db, 1)

err := pilot.Delete(db) // Regular variant, takes a db handle (boil.Executor interface).
pilot.DeleteP(db)       // Panic variant, takes a db handle and panics on error.
err := pilot.DeleteG()  // Global variant, uses the globally set db handle (boil.SetDB()).
pilot.DeleteGP()        // Global&Panic variant, combines the global db handle and panic on error.

Note that it's slightly different for query building.


Here are a list of all of the finishers that can be used in combination with Query Building.

Finishers all have P (panic) method variations. To specify your db handle use the G or regular variation of the Starter method.

// These are called like the following:

One() // Retrieve one row as object (same as LIMIT(1))
All() // Retrieve all rows as objects (same as SELECT * FROM)
Count() // Number of rows (same as COUNT(*))
UpdateAll(models.M{"name": "John", "age": 23}) // Update all rows matching the built query.
DeleteAll() // Delete all rows matching the built query.
Exists() // Returns a bool indicating whether the row(s) for the built query exists.
Bind(&myObj) // Bind the results of a query to your own struct object.
Raw Query

We provide boil.SQL() for executing raw queries. Generally you will want to use Bind() with this, like the following:

err := boil.SQL(db, "select * from pilots where id=$1", 5).Bind(&obj)

You can use your own structs or a generated struct as a parameter to Bind. Bind supports both a single object for single row queries and a slice of objects for multiple row queries.

You also have models.NewQuery() at your disposal if you would still like to use Query Build but would like to build against a non-generated model.


For a comprehensive ruleset for Bind() you can refer to our godoc.

The Bind() Finisher allows the results of a query built with the Raw SQL method or the Query Builder methods to be bound to your generated struct objects, or your own custom struct objects.

This can be useful for complex queries, queries that only require a small subset of data and have no need for the rest of the object variables, or custom join struct objects like the following:

// Custom struct using two generated structs
type PilotAndJet struct {
  models.Pilot `boil:",bind"`
  models.Jet   `boil:",bind"`

var paj PilotAndJet
// Use a raw query
err := boil.SQL(`
  select as "", as "", as "", jets.pilot_id as "jets.pilot_id",
  jets.age as "jets.age", as "", jets.color as "jets.color"
  from pilots inner join jets on jets.pilot_id=?`, 23,

// Use query building
err := models.NewQuery(db,
  Select("", "", "", "jets.pilot_id", "jets.age", "", "jets.color"),
  InnerJoin("jets on jets.pilot_id ="),
// Custom struct for selecting a subset of data
type JetInfo struct {
  AgeSum int `boil:"age_sum"`
  Count int `boil:"juicy_count"`

var info JetInfo

// Use query building
err := models.NewQuery(db, Select("sum(age) as age_sum", "count(*) as juicy_count", From("jets"))).Bind(&info)

// Use a raw query
err := boil.SQL(`select sum(age) as "age_sum", count(*) as "juicy_count" from jets`).Bind(&info)

We support the following struct tag modes for Bind() control:

type CoolObject struct {
  // Don't specify a name, Bind will TitleCase the column
  // name, and try to match against this.
  Frog int

  // Specify an alternative name for the column, it will
  // be titlecased for matching, can be whatever you like.
  Cat int  `boil:"kitten"`

  // Ignore this struct field, do not attempt to bind it.
  Pig int  `boil:"-"`

  // Instead of binding to this as a regular struct field
  // (like other sql-able structs eg. time.Time)
  // Recursively search inside the Dog struct for field names from the query.
  Dog      `boil:",bind"`

  // Same as the above, except specify a different table name
  Mouse    `boil:"rodent,bind"`

  // Ignore this struct field, do not attempt to bind it.
  Bird     `boil:"-"`

Helper methods will be generated for every to one and to many relationship structure you have defined in your database by using foreign keys.

We attach these helpers directly to your model struct, for example:

jet, _ := models.FindJet(db, 1)

// "to one" relationship helper method.
// This will retrieve the pilot for the jet.
pilot, err := jet.Pilot(db).One()

// "to many" relationship helper method.
// This will retrieve all languages for the pilot.
languages, err := pilot.Languages(db).All()

If your relationship involves a join table SQLBoiler will figure it out for you transparently.

It is important to note that you should use Eager Loading if you plan on loading large collections of rows, to avoid N+1 performance problems.

For example, take the following:

// Avoid this loop query pattern, it is slow.
jets, _ := models.Jets(db).All()
pilots := make([]models.Pilot, len(jets))
for i := 0; i < len(jets); i++ {
  pilots[i] = jets[i].Pilot(db).OneP()

// Instead, use Eager Loading!
jets, _ := models.Jets(db, Load("Pilot")).All()

Eager loading can be combined with other query mods, and it can also eager load recursively.

// Example of a nested load.
// Each jet will have its pilot loaded, and each pilot will have its languages loaded.
jets, _ := models.Jets(db, Load("Pilot.Languages")).All()
// Note that each level of a nested Load call will be loaded. No need to call Load() multiple times.

// A larger, random example
users, _ := models.Users(db,
  Where("age > ?", 23),

We provide the following methods for managing relationships on objects:

To One

  • SetX(): Set the foreign key to point to something else: jet.SetPilot(...)
  • RemoveX(): Null out the foreign key, effectively removing the relationship between these two objects: jet.RemovePilot(...)

To Many

  • AddX(): Add more relationships to the existing set of related Xs: pilot.AddLanguages(...)
  • SetX(): Remove all existing relationships, and replace them with the provided set: pilot.SetLanguages(...)
  • RemoveX(): Remove all provided relationships: pilot.RemoveLanguages(...)

To One code examples:

  jet, _ := models.FindJet(db, 1)
  pilot, _ := models.FindPilot(db, 1)

  // Set the pilot to an existing pilot
  err := jet.SetPilot(db, false, &pilot)

  pilot = models.Pilot{
    Name: "Erlich",

  // Insert the pilot into the database and assign it to a jet
  err := jet.SetPilot(db, true, &pilot)

  // Remove a relationship. This method only exists for foreign keys that can be NULL.
  err := jet.RemovePilot(db, &pilot)

To Many code examples:

  pilots, _ := models.Pilots(db).All()
  languages, _ := models.Languages(db).All()

  // Set a group of language relationships
  err := pilots.SetLanguages(db, false, &languages)

  languages := []*models.Language{
    {Language: "Strayan"},
    {Language: "Yupik"},
    {Language: "Pawnee"},

  // Insert new a group of languages and assign them to a pilot
  err := pilots.SetLanguages(db, true, languages...)

  // Add another language relationship to the existing set of relationships
  err := pilots.AddLanguages(db, false, &someOtherLanguage)

  anotherLanguage := models.Language{Language: "Archi"}

  // Insert and then add another language relationship
  err := pilots.AddLanguages(db, true, &anotherLanguage)

  // Remove a group of relationships
  err := pilots.RemoveLanguages(db, languages...)

Before and After hooks are available for most operations. If you don't need them you can shrink the size of the generated code by disabling them with the --no-hooks flag.

Every generated package that includes hooks has the following HookPoints defined:

const (
  BeforeInsertHook HookPoint = iota + 1

To register a hook for your model you will need to create the hook function, and attach it with the AddModelHook method. Here is an example of a before insert hook:

// Define my hook function
func myHook(exec boil.Executor, p *Pilot) {
  // Do stuff

// Register my before insert hook for pilots
models.AddPilotHook(boil.BeforeInsertHook, myHook)

Your ModelHook will always be defined as func(boil.Executor, *Model)


The boil.Executor interface powers all of SQLBoiler. This means anything that conforms to the three Exec/Query/QueryRow methods can be used. sql.DB, sql.Tx as well as other libraries (sqlx) conform to this interface, and therefore any of these things may be used as an executor for any query in the system. This makes using transactions very simple:

tx, err := db.Begin()
if err != nil {
  return err

users, _ := models.Pilots(tx).All()

// Rollback or commit
Debug Logging

Debug logging will print your generated SQL statement and the arguments it is using. Debug logging can be toggled on globally by setting the following global variable to true:

boil.DebugMode = true

// Optionally set the writer as well. Defaults to os.Stdout
fh, _ := os.Open("debug.txt")
boil.DebugWriter = fh

Note: Debug output is messy at the moment. This is something we would like addressed.


Select is done through Query Building and Find. Here's a short example:

// Select one pilot
pilot, err := models.Pilots(db, qm.Where("name=?", "Tim")).One()

// Select specific columns of many jets
jets, err := models.Jets(db, qm.Select("age", "name")).All()

Find is used to find a single row by primary key:

// Retrieve pilot with all columns filled
pilot, err := models.PilotFind(db, 1)

// Retrieve a subset of column values
jet, err := models.JetFind(db, 1, "name", "color")

The main thing to be aware of with Insert is how the whitelist operates. If no whitelist argument is provided, Insert will abide by the following rules:

  • Insert all columns without a database default value.
  • Insert all columns with a non-zero value that have a database default value.

On the other hand, if a whitelist is provided, we will only insert the columns specified in the whitelist.

Also note that your object will automatically be updated with any missing default values from the database after the Insert is finished executing. This includes auto-incrementing column values.

var p1 models.Pilot
p1.Name = "Larry"
err := p1.Insert(db) // Insert the first pilot with name "Larry"
// p1 now has an ID field set to 1

var p2 models.Pilot
p2.Name "Borris"
err := p2.Insert(db) // Insert the second pilot with name "Borris"
// p2 now has an ID field set to 2

var p3 models.Pilot
p3.ID = 25
p3.Name = "Rupert"
err := p3.Insert(db) // Insert the third pilot with a specific ID
// The id for this row was inserted as 25 in the database.

var p4 models.Pilot
p4.ID = 0
p4.Name = "Nigel"
err := p4.Insert(db, "id", "name") // Insert the fourth pilot with a zero value ID
// The id for this row was inserted as 0 in the database.
// Note: We had to use the whitelist for this, otherwise
// SQLBoiler would presume you wanted to auto-increment

Update can be performed on a single object, a slice of objects or as a Finisher for a collection of rows.

Update on a single object optionally takes a whitelist. The purpose of the whitelist is to specify which columns in your object should be updated in the database.

If no whitelist argument is provided, Update will update every column except for primary key columns.

If a whitelist argument is provided, update will only update the columns specified.

// Find a pilot and update his name
pilot, _ := models.FindPilot(db, 1)
pilot.Name = "Neo"
err := pilot.Update(db)

// Update a slice of pilots to have the name "Smith"
pilots, _ := models.Pilots(db).All()
err := pilots.UpdateAll(db, models.M{"name": "Smith"})

// Update all pilots in the database to to have the name "Smith"
err := models.Pilots(db).UpdateAll(models.M{"name", "Smith"})

Delete a single object, a slice of objects or specific objects through Query Building.

pilot, _ := models.FindPilot(db, 1)
// Delete the pilot from the database
err := pilot.Delete(db)

// Delete all pilots from the database
err := models.Pilots(db).DeleteAll()

// Delete a slice of pilots from the database
pilots, _ := models.Pilots(db).All()
err := pilots.DeleteAll(db)

Upsert allows you to perform an insert that optionally performs an update when a conflict is found against existing row values.

The whitelist operates in the same fashion that it does for Insert.

If an insert is performed, your object will be updated with any missing default values from the database, such as auto-incrementing column values.

var p1 models.Pilot
p1.ID = 5
p1.Name = "Gaben"

// INSERT INTO pilots ("id", "name") VALUES($1, $2)
err := p1.Upsert(db, false, nil, nil)

// INSERT INTO pilots ("id", "name") VALUES ($1, $2)
// ON CONFLICT ("id") DO UPDATE SET "name" = EXCLUDED."name"
err := p1.Upsert(db, true, []string{"id"}, []string{"name"})

// Set p1.ID to a zero value. We will have to use the whitelist now.
p1.ID = 0
p1.Name = "Hogan"

// INSERT INTO pilots ("id", "name") VALUES ($1, $2)
// ON CONFLICT ("id") DO UPDATE SET "name" = EXCLUDED."name"
err := p1.Upsert(db, true, []string{"id"}, []string{"name"}, "id", "name")

Note: Passing a different set of column values to the update component is not currently supported. If this feature is important to you let us know and we can consider adding something for this.


In the event that your objects get out of sync with the database for whatever reason, you can use Reload and ReloadAll to reload the objects using the primary key values attached to the objects.

pilot, _ := models.FindPilot(db, 1)

// > Object becomes out of sync for some reason

// Refresh the object with the latest data from the db
err := pilot.Reload(db)

// Reload all objects in a slice
pilots, _ := models.Pilots(db).All()
err := pilots.ReloadAll(db)

Note: Reload and ReloadAll are not recursive, if you need your relationships reloaded you will need to call the Reload methods on those yourself.

jet, err := models.FindJet(db, 1)

// Check if the pilot assigned to this jet exists.
exists := jet.Pilot(db).Exists()

// Check if the pilot with ID 5 exists
exists := models.Pilots(db, Where("id=?", 5)).Exists()


Won't compiling models for a huge database be very slow?

No, because Go's toolchain - unlike traditional toolchains - makes the compiler do most of the work instead of the linker. This means that when the first go install is done it can take a little bit of time because there is a lot of code that is generated. However, because of this work balance between the compiler and linker in Go, linking to that code afterwards in the subsequent compiles is extremely fast.

Missing imports for generated package

The generated models might import a couple of packages that are not on your system already, so cd into your generated models directory and type go get -u -t to fetch them. You will only need to run this command once, not per generation.


If you'd like to run the benchmarks yourself check out our boilbench repo.

Here are the results (lower is better):

go test -bench . -benchmem

BenchmarkGORMDelete/gorm-8     	    100000	     15364 ns/op	    5395 B/op	     113 allocs/op
BenchmarkGORPDelete/gorp-8     	   1000000	      1703 ns/op	     304 B/op	      12 allocs/op
BenchmarkXORMDelete/xorm-8     	    100000	     14733 ns/op	    3634 B/op	     107 allocs/op
BenchmarkBoilDelete/boil-8     	   2000000	       986 ns/op	     120 B/op	       7 allocs/op

BenchmarkGORMInsert/gorm-8     	    100000	     19197 ns/op	    8054 B/op	     161 allocs/op
BenchmarkGORPInsert/gorp-8     	    500000	      3413 ns/op	    1008 B/op	      32 allocs/op
BenchmarkXORMInsert/xorm-8     	    100000	     15428 ns/op	    5836 B/op	     131 allocs/op
BenchmarkBoilInsert/boil-8     	    500000	      3041 ns/op	     568 B/op	      21 allocs/op

BenchmarkGORMSelectAll/gorm-8  	     20000	     85422 ns/op	   29912 B/op	     511 allocs/op
BenchmarkGORPSelectAll/gorp-8  	     50000	     35824 ns/op	    8837 B/op	     312 allocs/op
BenchmarkXORMSelectAll/xorm-8  	     30000	     58843 ns/op	   13805 B/op	     298 allocs/op
BenchmarkBoilSelectAll/boil-8  	    100000	     13844 ns/op	    2840 B/op	      61 allocs/op

BenchmarkGORMSelectSubset/gorm-8     10000	    100714 ns/op	   30875 B/op	     517 allocs/op
BenchmarkGORPSelectSubset/gorp-8     30000	     43547 ns/op	    8837 B/op	     312 allocs/op
BenchmarkXORMSelectSubset/xorm-8     30000	     48128 ns/op	   12989 B/op	     282 allocs/op
BenchmarkBoilSelectSubset/boil-8    100000	     12316 ns/op	    2977 B/op	      65 allocs/op

BenchmarkGORMSelectComplex/gorm-8    10000	    133598 ns/op	   49398 B/op	     772 allocs/op
BenchmarkGORPSelectComplex/gorp-8    50000	     40588 ns/op	    9037 B/op	     321 allocs/op
BenchmarkXORMSelectComplex/xorm-8    30000	     56367 ns/op	   14174 B/op	     313 allocs/op
BenchmarkBoilSelectComplex/boil-8   100000	     16941 ns/op	    3821 B/op	      95 allocs/op

BenchmarkGORMUpdate/gorm-8           50000	     25406 ns/op	    9710 B/op	     195 allocs/op
BenchmarkGORPUpdate/gorp-8          300000	      3614 ns/op	    1152 B/op	      34 allocs/op
BenchmarkXORMUpdate/xorm-8          100000	     17510 ns/op	    4458 B/op	     132 allocs/op
BenchmarkBoilUpdate/boil-8          500000	      2958 ns/op	     520 B/op	      16 allocs/op

BenchmarkGORMRawBind/gorm-8    	     10000	    112577 ns/op	   38270 B/op	     595 allocs/op
BenchmarkGORPRawBind/gorp-8    	     30000	     40967 ns/op	    8837 B/op	     312 allocs/op
BenchmarkXORMRawBind/xorm-8    	     30000	     54739 ns/op	   12692 B/op	     273 allocs/op
BenchmarkSQLXRawBind/sqlx-8    	    200000	     13537 ns/op	    4268 B/op	      49 allocs/op
BenchmarkBoilRawBind/boil-8    	    200000	     11144 ns/op	    4334 B/op	      49 allocs/op
Expand ▾ Collapse ▴



    Package main defines a command line interface for the sqlboiler package

    Package sqlboiler has types and methods useful for generating code that acts as a fully dynamic ORM might.


    Path Synopsis
    Package bdb supplies the sql(b)oiler (d)ata(b)ase abstractions.
    Package bdb supplies the sql(b)oiler (d)ata(b)ase abstractions.
    Package randomize has helpers for randomization of structs and fields
    Package randomize has helpers for randomization of structs and fields
    Package strmangle is a collection of string manipulation functions.
    Package strmangle is a collection of string manipulation functions.