Beam: a type-safe, non-TH Haskell relational database library and ORM

If you use beam commercially, please consider a donation to make this project possible: https://liberapay.com/tathougies

Beam is a Haskell interface to relational databases. Beam uses the Haskell type system to verify that queries are type-safe before sending them to the database server. Queries are written in a straightforward, natural monadic syntax. Combinators are provided for all standard SQL92 features, and a significant subset of SQL99, SQL2003, and SQL2008 features. For your convenience a thorough compatibility matrix is maintained here.

Beam is standards compliant but not naive. We recognize that different database backends provide different guarantees, syntaxes, and advantages. To reflect this, beam maintains a modular design. While the core package provides standard functionality, beam is split up into a variety of backends which provide a means to interface Beam's data query and update DSLs with particular RDBMS backends. Backends can be written and maintained independently of this repository. For example, the beam-mysql and beam-firebird backends are packaged independently.

Recognizing that over-abstraction frequently means caving in to the lowest common denominator, Beam does not do connection or transaction management. Rather, the user is free to perform these functions using the appropriate Haskell interface library for their backend of choice. Additionally, beam backends provide a significant portion of backend-specific functionality which seamlessly fits into the beam ecosystem.

For example, the beam-postgres backend is built off of the postgresql-simple interface library. When using beam-postgres, the user manages connections and transactions with postgresql-simple. The user is free to issue queries directly with postgresql-simple, only using beam when desired. Postgres offers a number of rich data types on top of the standard SQL data types. To reflect this, beam-postgres offers pluggable support for postgres-specific data types and features.

For more information, see the user guide.

For questions, feel free to join our mailing list or head over to #haskell-beam on freenode.

A word on testing

beam-core has in-depth unit tests to test query generation over an idealized ANSI SQL-compliant backend. You may be concerned that there are no tests in either beam-sqlite or beam-postgres. Do not be alarmed. The documentation contains many, many examples of queries written over the sample Chinook database, the schema for which can be found at beam-sqlite/examples/Chinook/Schema.hs. The included mkdocs configuration and custom beam_query python Markdown extension automatically run every query in the documentation against a live database connection. Any errors in serializion/deserialization or invalid syntax are caught while building the documentation. Feel free to open pull-requests with additional examples/tests.

Tests are written

!beam-query
```haskell
!example <template-name> <requirements>
do x <- all_ (customer chinookDb) -- chinookDb available under chinook and chinookdml examples
   pure x
```

The !beam-query declaration indicates this is markdown code block that contains beam query code. The !example declaration indicates that this example should be built against applicable backends and included in the code. The template_name is either chinook or chinookdml (depending on whether you have quest a query or a DML statement). For chinook, the included code should produce a Q query. For chinookdml, the included code should be a monadic action in a MonadBeam. The requirements can be used to select which backends to run this against. See the documentation for examples.

Building the documentation

Beam uses mkdocs for its documentation generation. The included build-docs.sh script can take care of building the documentation and serving it locally. In order to use the tool though, make sure you have a python installation with the mkdocs module installed. You can do this by creating a virtualenv, and pip installing mkdocs, or in a Nix shell with nix-shell docs.

The documentation uses a custom Markdown preprocessor to automatically build examples against the canonical Chinook database. By default, beam will build examples for every beam backend it knows about, including ones not in the main source tree (see docs/beam.yaml for the full configuration). This means you will need to have an instance of all these database servers running and available. This is usually not what you want.

To only build examples for a particular backend, modify mkdocs.yaml and set the enabled_backends configuration setting for the docs.markdown.beam_query preprocessor. For example, to only build docs for beam-sqlite, change

  - docs.markdown.beam_query:
      template_dir: 'docs/beam-templates'
      cache_dir: 'docs/.beam-query-cache'
      conf: 'docs/beam.yaml'
      base_dir: '.'

to

  - docs.markdown.beam_query:
      template_dir: 'docs/beam-templates'
      cache_dir: 'docs/.beam-query-cache'
      conf: 'docs/beam.yaml'
      base_dir: '.'
      enabled_backends:
        - beam-sqlite