Postgrest GO

[!IMPORTANT] Update, September 2026:
This codebase is not being actively maintained by Supabase or the community, but it's not the end of the road... We're busy building an official Supabase SDK for Go that will make this repository and the module it publishes obsolete. This thread has more information, as well as providing a venue for discussion in the meantime.

Golang client for PostgREST. The goal of this library is to make an "ORM-like" restful interface.

Documentation

Full documentation can be found here.

Installation

go get github.com/supabase-community/postgrest-go

Version compatibility

go get without a version installs the latest tagged release (currently v0.0.12). The main branch has a newer API with generics, structured responses, and context-aware Execute methods.

Install command	API
go get github.com/supabase-community/postgrest-go	v0.0.12 — Select(columns, count, head), Execute() returns ([]byte, int64, error)
go get github.com/supabase-community/postgrest-go@main	main — Select(columns, opts), Execute(ctx) returns *PostgrestResponse

The examples below (except Basic usage v0.0.12) require the main branch:

go get github.com/supabase-community/postgrest-go@main

Quick Start

Basic usage (v0.0.12) {#basic-usage-v012}

If you installed the default release tag, use this API:

package main

import (
	"encoding/json"
	"fmt"

	"github.com/supabase-community/postgrest-go"
)

func main() {
	client := postgrest.NewClient("http://localhost:3000/rest/v1", "public", nil)
	if client.ClientError != nil {
		panic(client.ClientError)
	}

	data, count, err := client.From("users").Select("*", "", false).Execute()
	if err != nil {
		panic(err)
	}

	fmt.Printf("count: %d\n", count)
	fmt.Println(string(data))
}

Basic Usage (main branch)

package main

import (
	"context"
	"encoding/json"
	"fmt"

	"github.com/supabase-community/postgrest-go"
)

func main() {
	// Create a new client
	client := postgrest.NewClient("http://localhost:3000/rest/v1", "public", nil)
	if client.ClientError != nil {
		panic(client.ClientError)
	}

	// Select data
	response, err := client.From("users").Select("*", nil).Execute(context.Background())
	if err != nil {
		panic(err)
	}

	if response.Error != nil {
		panic(response.Error)
	}

	// Print results
	data, _ := json.Marshal(response.Data)
	fmt.Println(string(data))
}

Select with Filters

// Select with filters
opts := &postgrest.SelectOptions{Count: "exact"}
response, err := client.
	From("users").
	Select("id, name, email", opts).
	Eq("status", "active").
	Limit(10, nil).
	Order("name", &postgrest.OrderOptions{Ascending: true}).
	Execute(context.Background())

Insert Data

// Insert a single row
insertOpts := &postgrest.InsertOptions{Count: "exact"}
response, err := client.
	From("users").
	Insert(map[string]interface{}{
		"name":  "John Doe",
		"email": "[email protected]",
	}, insertOpts).
	Execute(context.Background())

Update Data

// Update rows
updateOpts := &postgrest.UpdateOptions{Count: "exact"}
response, err := client.
	From("users").
	Update(map[string]interface{}{
		"status": "inactive",
	}, updateOpts).
	Eq("id", 1).
	Execute(context.Background())

Delete Data

// Delete rows
deleteOpts := &postgrest.DeleteOptions{Count: "exact"}
response, err := client.
	From("users").
	Delete(deleteOpts).
	Eq("status", "inactive").
	Execute(context.Background())

Upsert Data

// Upsert (insert or update)
upsertOpts := &postgrest.UpsertOptions{
	OnConflict:      "email",
	IgnoreDuplicates: false,
	Count:           "exact",
}
response, err := client.
	From("users").
	Upsert(map[string]interface{}{
		"email": "[email protected]",
		"name":  "John Doe",
	}, upsertOpts).
	Execute(context.Background())

RPC (Remote Procedure Call)

// Call a PostgreSQL function
rpcOpts := &postgrest.RpcOptions{Count: "exact"}
response, err := client.
	Rpc("get_user_by_email", map[string]interface{}{
		"email": "[email protected]",
	}, rpcOpts).
	Execute(context.Background())

Advanced Filtering

// Multiple filters
response, err := client.
	From("users").
	Select("*", nil).
	Eq("status", "active").
	Gte("age", 18).
	Lte("age", 65).
	In("role", []interface{}{"admin", "user"}).
	Like("name", "%John%").
	Execute(context.Background())

// Text search
textSearchOpts := &postgrest.TextSearchOptions{
	Type:   "websearch",
	Config: "english",
}
response, err := client.
	From("posts").
	Select("*", nil).
	TextSearch("content", "golang tutorial", textSearchOpts).
	Execute(context.Background())

Single Result

// Get a single row
response, err := client.
	From("users").
	Select("*", nil).
	Eq("id", 1).
	Limit(1, nil).
	Single().
	Execute(context.Background())

// Get a single row or null (maybeSingle)
response, err := client.
	From("users").
	Select("*", nil).
	Eq("id", 999).
	MaybeSingle().
	Execute(context.Background())

ExecuteTo (Unmarshal directly)

// Unmarshal directly into a struct
type User struct {
	ID    int64  `json:"id"`
	Name  string `json:"name"`
	Email string `json:"email"`
}

var users []User
count, err := client.
	From("users").
	Select("*", nil).
	Eq("status", "active").
	ExecuteTo(context.Background(), &users)

Schema Selection

// Switch to a different schema
client = client.Schema("private")
response, err := client.From("sensitive_data").Select("*", nil).Execute(context.Background())

Error Handling

response, err := client.From("users").Select("*", nil).Execute(context.Background())
if err != nil {
	// Network or parsing error
	panic(err)
}

if response.Error != nil {
	// PostgREST error
	fmt.Printf("Error: %s (Code: %s)\n", response.Error.Message, response.Error.Code)
	fmt.Printf("Details: %s\n", response.Error.Details)
	fmt.Printf("Hint: %s\n", response.Error.Hint)
}

Throw on Error

// Throw errors instead of returning them in response
response, err := client.
	From("users").
	Select("*", nil).
	ThrowOnError().
	Execute(context.Background())
// If there's an error, it will be thrown here

API Reference

Client Methods

• NewClient(url, schema, headers) - Create a new client
• From(table) - Start a query on a table
• Rpc(function, args, opts) - Call a PostgreSQL function
• Schema(schema) - Switch to a different schema
• SetApiKey(key) - Set API key header
• SetAuthToken(token) - Set authorization token
• ChangeSchema(schema) - Change schema for subsequent requests

QueryBuilder Methods

• Select(columns, opts) - Select columns
• Insert(values, opts) - Insert rows
• Update(values, opts) - Update rows
• Upsert(values, opts) - Upsert rows
• Delete(opts) - Delete rows

FilterBuilder Methods

• Eq(column, value) - Equal
• Neq(column, value) - Not equal
• Gt(column, value) - Greater than
• Gte(column, value) - Greater than or equal
• Lt(column, value) - Less than
• Lte(column, value) - Less than or equal
• Like(column, pattern) - Case-sensitive pattern match
• Ilike(column, pattern) - Case-insensitive pattern match
• Is(column, value) - IS operator (for NULL checks)
• In(column, values) - IN operator
• Contains(column, value) - Contains (for arrays/jsonb)
• ContainedBy(column, value) - Contained by
• TextSearch(column, query, opts) - Full-text search
• Match(query) - Match multiple columns
• Not(column, operator, value) - Negate operator
• Or(filters, opts) - OR condition

TransformBuilder Methods

• Order(column, opts) - Order results
• Limit(count, opts) - Limit results
• Range(from, to, opts) - Range results
• Single() - Get single result
• MaybeSingle() - Get single result or null
• Select(columns) - Select after insert/update/delete
• CSV() - Return CSV format
• GeoJSON() - Return GeoJSON format
• Explain(opts) - Get query plan
• Rollback() - Rollback transaction
• MaxAffected(value) - Set max affected rows

Response Structure

type PostgrestResponse[T any] struct {
	Error      *PostgrestError `json:"error,omitempty"`
	Data       T               `json:"data,omitempty"`
	Count      *int64          `json:"count,omitempty"`
	Status     int             `json:"status"`
	StatusText string          `json:"statusText"`
}

Testing

Unit Tests

Some tests are implemented to run against mocked Postgrest endpoints using httpmock. To run unit tests:

go test ./...

Integration Tests

Integration tests use testcontainers-go to spin up real PostgreSQL and PostgREST containers for testing. This ensures tests run against actual PostgREST instances.

Running Integration Tests

Option 1: Using Testcontainers (Default)

Integration tests will automatically use testcontainers to start PostgreSQL and PostgREST containers:

go test -v -run TestIntegration

Option 2: Using Existing Services

If you have a running PostgREST instance, you can use it instead:

export POSTGREST_URL=http://localhost:3000
export USE_TESTCONTAINERS=false
go test -v -run TestIntegration

Option 3: Using Docker Compose

You can also use the provided docker-compose.yaml to start services manually:

docker-compose up -d
export POSTGREST_URL=http://localhost:3000
export USE_TESTCONTAINERS=false
go test -v -run TestIntegration
docker-compose down

Integration Test Coverage

The integration tests cover:

• ✅ Basic SELECT queries
• ✅ Filtering (Eq, In, Like, etc.)
• ✅ Ordering and limiting
• ✅ Single and MaybeSingle results
• ✅ INSERT operations
• ✅ UPDATE operations
• ✅ DELETE operations
• ✅ RPC (Remote Procedure Calls)
• ✅ Schema switching
• ✅ Complex queries with multiple filters
• ✅ Error handling
• ✅ Relationships and joins

Test Database Schema

The integration tests use the schema defined in test/00-schema.sql and seed data from test/01-dummy-data.sql. These files define:

• users table with various data types
• channels and messages tables for relationship testing
• kitchen_sink table for comprehensive type testing
• movie, person, profile tables for foreign key testing
• PostgreSQL functions for RPC testing
• Multiple schemas (public and personal) for schema switching tests

Package made possible through the efforts of:

Made with contrib.rocks.

License

This repo is licensed under the Apache License.

Sponsors

We are building the features of Firebase using enterprise-grade, open source products. We support existing communities wherever possible, and if the products don't exist we build them and open source them ourselves. Thanks to these sponsors who are making the OSS ecosystem better for everyone.