-
Notifications
You must be signed in to change notification settings - Fork 122
Model
go-jet edited this page Nov 3, 2024
·
10 revisions
Model file contains simple Go struct type used to store result of SQL queries. Model types can be used alone or combined to form complex object composition to store database query result. They are auto-generated from database tables, views and enums.
Following rules are applied to generate model types from database tables and views:
- for every table there is one Go file generated. File name is in snake case of the table name.
- every model file contains one struct type. Type name is a camel case of table name. Package name
is always
model
. - for every column of table there is a field in model struct type. Field name is camel case of column name. See below table for type mapping.
- fields are pointer types, if they relate to column that can be NULL.
- fields corresponds to primary key columns are tagged with
sql:"primary_key"
. This tag is used during query execution to group row results into desired arbitrary structure. See more at Query Result Mapping (QRM)
Database type(postgres) | Go type |
---|---|
boolean | bool |
smallint | int16 |
integer | int32 |
bigint | int64 |
real | float32 |
numeric, decimal, double precision | float64 |
date, timestamp, time(with or without timezone) | time.Time |
bytea | []byte |
uuid | uuid.UUID |
enum | enum name |
text, character, character varying, | |
and all remaining types | string |
Database type(postgres) | Go type |
---|---|
boolean or BIT(1) | bool |
tinylint [unsigned] | [u]int8 |
smallint [unsigned] | [u]int16 |
mediumint [unsigned] | [u]int32 |
integer [unsigned] | [u]int32 |
bigint [unsigned] | [u]int64 |
real | float32 |
numeric, decimal, double precision | float64 |
date, time, datetime, timestamp | time.Time |
binary, varbinary, tinyblob, blob, | |
mediumblob, longblob | []byte |
enum | table name + enum name |
text, character, character varying, | |
and all remaining types | string |
PostgreSQL table address
:
CREATE TABLE dvds.address
(
address_id serial NOT NULL DEFAULT,
address character varying(50) NOT NULL,
address2 character varying(50),
district character varying(20) NOT NULL,
city_id smallint NOT NULL,
postal_code character varying(10),
phone character varying(20) NOT NULL,
last_update timestamp without time zone NOT NULL DEFAULT now(),
CONSTRAINT address_pkey PRIMARY KEY (address_id)
)
Autogenerated model file address.go
:
package model
import (
"time"
)
type Address struct {
AddressID int32 `sql:"primary_key"`
Address string
Address2 *string
District string
CityID int16
PostalCode *string
Phone string
LastUpdate time.Time
}
Following rules are applied to generate model files from database enums:
- for every enum there is one Go file generated.
- PostgreSQL: File name is a snake case of
enum name
. - MySQL or MariaDB: File name is snake case of
table name
+enum name
.
- PostgreSQL: File name is a snake case of
- every file contains one renamed string type. Type name is a camel case of enum name.
Package name is always
model
. Enum type has two helper methods to:- initialize correctly from database query result
- easily convert enum to string.
- for every enum value there is one constant defined.
Name of the constant is in format
{CamelCase(enum_name)}_{CamelCase(enum_value_name)}
. - there is one slice contains list of all enums defined named
{CamelCase(enum_name)}AllValues
PostgreSQL:
CREATE TYPE dvds.mpaa_rating AS ENUM
('G', 'PG', 'PG-13', 'R', 'NC-17');
Autogenerated model file mpaa_rating.go
package model
import "errors"
type MpaaRating string
const (
MpaaRating_G MpaaRating = "G"
MpaaRating_Pg MpaaRating = "PG"
MpaaRating_Pg13 MpaaRating = "PG-13"
MpaaRating_R MpaaRating = "R"
MpaaRating_Nc17 MpaaRating = "NC-17"
)
var MpaaRatingAllValues = []MpaaRating{
MpaaRating_G,
MpaaRating_Pg,
MpaaRating_Pg13,
MpaaRating_R,
MpaaRating_Nc17,
}
func (e *MpaaRating) Scan(value interface{}) error {
var enumValue string
switch val := value.(type) {
case string:
enumValue = val
case []byte:
enumValue = string(val)
default:
return errors.New("jet: Invalid scan value for AllTypesEnum enum. Enum value has to be of type string or []byte")
}
switch enumValue {
case "G":
*e = MpaaRating_G
case "PG":
*e = MpaaRating_Pg
case "PG-13":
*e = MpaaRating_Pg13
case "R":
*e = MpaaRating_R
case "NC-17":
*e = MpaaRating_Nc17
default:
return errors.New("jet: Invalid scan value '" + enumValue + "' for MpaaRating enum")
}
return nil
}
func (e MpaaRating) String() string {
return string(e)
}
MySQL or MariaDB:
CREATE TABLE film (
rating ENUM('G','PG','PG-13','R','NC-17') DEFAULT 'G'
)
package model
import "errors"
type FilmRating string
const (
FilmRating_G FilmRating = "G"
FilmRating_Pg FilmRating = "PG"
FilmRating_Pg13 FilmRating = "PG-13"
FilmRating_R FilmRating = "R"
FilmRating_Nc17 FilmRating = "NC-17"
)
func (e *FilmRating) Scan(value interface{}) error {
if v, ok := value.(string); !ok {
return errors.New("jet: Invalid data for FilmRating enum")
} else {
switch string(v) {
case "G":
*e = FilmRating_G
case "PG":
*e = FilmRating_Pg
case "PG-13":
*e = FilmRating_Pg13
case "R":
*e = FilmRating_R
case "NC-17":
*e = FilmRating_Nc17
default:
return errors.New("jet: Inavlid data " + string(v) + "for FilmRating enum")
}
return nil
}
}
func (e FilmRating) String() string {
return string(e)
}
- Home
- Generator
- Model
- SQL Builder
- Query Result Mapping (QRM)
-
FAQ
- How to execute jet statement in SQL transaction?
- How to construct dynamic projection list?
- How to construct dynamic condition?
- How to use jet in multi-tenant environment?
- How to change model field type?
- How to use custom(or currently unsupported) functions and operators?
- How to use IN/NOT_IN with dynamic list of values?
- Scan stopped working after naming a destination type