Skip to the content
Know your tech before you buy it. Makers: get listed from $49/mo

pgloader

pgloader.io

  • Free plan
  • Linux
  • Self-hosted
7.2 Club score Visit pgloader.io

What pgloader is

pgloader loads database and flat-file data into PostgreSQL, translating schemas and applying casting rules during migration. Listed sources include MySQL, MariaDB, MS SQL Server, SQLite, CSV, fixed-width files, and DBF; Oracle support is not built yet. Its pipeline fetches the source catalog, prepares the target, runs parallel COPY, creates indexes in parallel, and completes the schema. MySQL handling includes enum translation, TINYINT(1)-to-boolean conversion, zero-date handling, and .my.cnf credential lookup. The SQL Server path supports Azure AD authentication, sequence migration, and multi-column foreign keys. For file imports, failed COPY transactions can be retried by isolating rejected rows while accepted rows continue loading. Custom casting rules can be set in a .load file. The stable packaged release is v3 in Common Lisp; v4 is a pre-release self-contained JAR requiring Java 21 or later. It is self-hosted on Linux. The free Community tier offers best-effort GitHub help; paid support starts at €100/mo.

Who it suits

It suits teams migrating supported database or flat-file sources into PostgreSQL who can run a self-hosted Linux tool. Users needing priority help can consider the paid support tiers.

Good to know, and what to watch

  • Loads data using parallel COPY and index creation.
  • Supports a range of database and flat-file sources.
  • Custom casting rules can be set in a .load file.
  • Can continue file imports while separating rejected rows.
  • Oracle support is not built yet.
  • Version 4 is a pre-release requiring Java 21 or later.
  • The free Community tier offers best-effort GitHub help.
  • It does not currently parse SQL object definitions.

Community

Free

Supporter

€100/mo

Professional

€800/mo

Sponsor

€2,000/mo

Free plan

Getting started

  1. Open the pgloader website
  2. Prepare a self-hosted Linux environment
  3. Choose stable packaged v3 or the v4 pre-release JAR
  4. For v4, install Java 21 or later
  5. Define any custom casting rules in a .load file
  6. Select the Community tier or a paid support plan

Limits to check

The free Community plan provides best-effort GitHub support. Version 4 requires Java 21 or later and is a pre-release; catalog-based operation does not currently translate view definitions or stored procedures.

The Club's review

pgloader is a self-hosted utility for moving supported databases and flat files into PostgreSQL. It fits engineers and administrators who can run their own migration jobs and need source-aware conversion rather than a hosted service. The strongest case is a controlled, one-time or staged move where parallel loading and custom casting save manual work.

Overview

pgloader structures a migration as catalog retrieval, target preparation, parallel COPY loading, parallel index creation, and schema completion. That staged pipeline gives teams a clear path from source inspection to a populated PostgreSQL database, with parallel work suited to substantial datasets. It also provides discovery and assessment and dependency mapping, but it does not provide continuous replication, so it is not the right tool for keeping two live systems synchronized.

Sources include MySQL, MariaDB, Microsoft SQL Server, SQLite, CSV, fixed-width files, and DBF. Oracle is not supported yet, which is a decisive limitation for Oracle estates. Dimitri Fontaine has authored and maintained the project since 2005. Deployment is self-hosted on Linux, keeping operation under your control but leaving installation, runtime management, and migration execution to your team.

Key features

Source-aware conversion

The MySQL route handles enum-to-type translation, converts TINYINT(1) values to booleans, deals with zero dates, and can read credentials from .my.cnf. Those are practical fixes for common MySQL-to-PostgreSQL friction and reduce cleanup after a load. The SQL Server route adds Azure AD authentication, sequence migration, and multi-column foreign-key support, giving SQL Server teams coverage for relationships and authentication patterns that basic bulk-copy tools may leave to manual work.

Custom rules and resilient file loading

Teams can define custom casting rules in a .load file when the standard conversions do not match application requirements. File imports are designed to keep useful data moving: when a COPY transaction fails, rejected rows are separated and the accepted rows can still load after a retry. That behavior is valuable for imperfect CSV or fixed-width inputs, although rejected records still require review before you can call the migration complete.

Runtime and object limits

Version 4 is distributed as a self-contained JAR and requires Java 21 or later. The project describes v4 as pre-release, while the Common Lisp v3 package is the stable release, so production teams should choose their version deliberately. The roadmap also says pgloader works from catalog representations and does not currently parse SQL object definitions. View definitions and stored procedures therefore need separate SQL translation work; pgloader should be treated as a data-and-schema migration component, not a complete application-database refactoring system.

Pricing

The Community plan is free (free) for evaluation or non-critical use, with best-effort support through GitHub. It is the sensible starting point for trials, internal projects, and teams able to diagnose migration issues themselves.

Supporter costs €100 a month, billed €100/mo, and adds a priority queue with answers within a few days. Professional costs €800 a month, billed €800/mo, with patches triaged under an SLA and targeted for the next release. Sponsor costs €2,000 a month, billed €2,000/mo, adding roadmap input and reserved development bandwidth. These paid tiers suit organisations that need maintainer attention or influence over delivery rather than extra hosted capacity.

For a single urgent blocker, Fast-Lane is €3,000, paid once, billed €3,000 per issue, with guaranteed resolution in the next release and no subscription. Per-Release Sponsor is €10,000, paid once, billed €10,000 per release, for prioritised pull requests and features, release-note credit, and one sponsor per release. The paid options buy support and development priority; they do not turn pgloader into a managed migration service.

Platforms

pgloader runs on Linux and uses a self-hosted deployment model. Version 4's Java 21-or-later requirement is relevant to build and operations planning; teams standardising on older Java runtimes may prefer the stable packaged v3 route.

Who it's for

Choose pgloader when you are moving MySQL, MariaDB, SQL Server, SQLite, or supported file formats into PostgreSQL and want parallel loading, built-in source conversions, and configurable casting under your own control. It is particularly appropriate for Linux-based teams with database expertise and for migrations where discovery, dependency mapping, and repeatable load files matter.

Look elsewhere for Oracle migrations, ongoing replication, or projects that must automatically carry views and stored procedures across. The pre-release status of v4 also makes release selection and operational risk an explicit part of the decision.

Pros and cons

  • Pros: Parallel COPY and index creation can shorten staged migrations.
  • Pros: MySQL and SQL Server conversions address concrete type, date, authentication, sequence, and key-relationship issues.
  • Pros: Custom .load casting and rejected-row handling give operators control over messy inputs.
  • Cons: Oracle is not a supported source.
  • Cons: No continuous replication, so it cannot serve as a live synchronization layer.
  • Cons: SQL object definitions such as views and stored procedures need separate translation.
  • Cons: v4 is pre-release and requires Java 21 or later.

Alternatives

For a broad survey of migration products, see Cloud Migration Tools.

Google Compute Engine is a freemium API, Linux, web, and Windows platform with a free tier, so it fits teams that want infrastructure on which to run migration workloads rather than a dedicated PostgreSQL migration utility.

Striim offers freemium, self-hosted and web deployment with a Developer plan capped at 25 million events per month; choose it when event-based data pipelines are the priority.

Google Cloud Migration Center is free for assessments and cost estimates, making it a better starting point for discovery than for executing pgloader-style data loads.

Amazon Database Migration Service is a paid API and web service with a free tier and credits, suited to teams seeking a managed AWS migration service.

CAST Highlight is a paid application-assessment product with a free trial, useful when application insight is the primary requirement.

Tencent Cloud Cloud Migration Service is free to use while migration resources and ongoing services are charged separately, fitting Tencent Cloud-oriented moves.

AWS Transform MGN (formerly AWS Application Migration Service) provides a 90-day free period per source server, making it relevant to server-replication projects rather than database conversion into PostgreSQL.

Device42 is a paid, trial-available platform with custom annual pricing based on device count, better suited to infrastructure discovery and inventory.

Verdict

pgloader is a strong choice for technically capable Linux teams migrating supported sources into PostgreSQL with parallel loading and adjustable casting rules. Its focused conversions and self-hosted control are the main reasons to choose it. Look elsewhere if you need Oracle coverage, continuous replication, automatic migration of views and stored procedures, or a fully managed service.

The verdict

pgloader offers a staged route into PostgreSQL for its listed sources, with options for custom casting and parallel loading. Check source support and release status, especially if you depend on Oracle or need SQL object definitions migrated.

Questions people ask

How much does pgloader cost?

The Community plan is free. Paid support starts with Supporter at €100 a month; other listed tiers include Professional at €800 a month and Sponsor at €2,000 a month.

Which sources can pgloader migrate?

Listed sources are MySQL, MariaDB, MS SQL Server, SQLite, CSV, fixed-width files, and DBF.

What platforms does pgloader support?

pgloader is self-hosted on Linux.

Which release should I use?

Version 3 in Common Lisp is the stable packaged release. Version 4 is a pre-release self-contained JAR requiring Java 21 or later.

Does pgloader support continuous replication?

No. The listed capability is database migration rather than continuous replication.

Who maintains pgloader?

Dimitri Fontaine is identified as its author and maintainer since 2005.

What the maker publishes

Custom rules
Custom casting rules can be specified in a .load file. source
File imports
For file sources, failed COPY transactions are retried by separating rejected rows while allowing accepted rows to load. source
Founded
2005 source
Limits
The roadmap says pgloader works from catalog representations and does not currently parse SQL object definitions; view definitions and stored procedures would require SQL translation support. source
Maker
The maker identifies Dimitri Fontaine as pgloader’s author and maintainer since 2005. source
Migration sources
The site lists MySQL, MariaDB, MS SQL Server, SQLite, CSV, fixed-width, and DBF sources; Oracle support is not built yet. source
MySQL features
The MySQL path includes enum-to-type translation, TINYINT(1)-to-boolean conversion, zero-date handling, and .my.cnf credential lookup. source
Pipeline
Its five stages are catalog fetch, target preparation, parallel COPY, parallel index creation, and schema completion. source
Purpose
pgloader migrates data into PostgreSQL from databases and flat files using schema translation, casting rules, and parallel loading. source
Runtime
Version 4 is distributed as a self-contained JAR and requires Java 21 or later. source
Security
The site mentions Azure AD authentication for MS SQL Server but does not state security certifications or compliance attestations. source
SQL Server features
The MS SQL Server path includes Azure AD authentication, sequence migration, and multi-column foreign key support. source
Support
The free Community tier receives best-effort help through GitHub, while paid tiers offer priority handling and related support benefits. source
Version status
The site describes v4 as a pre-release and v3 in Common Lisp as the stable, packaged release. source

Checked against the maker's pages on 2 Oct 2026.