Installing SDF

SDF is installed as a single, dependency free binary on a host machine. The install script will first check your operating system and architecture, then install the appropriate latest stable version of SDF and its dependencies for your system.

The sdf command will also be added to your path.

To the latest stable release of SDF:

curl -LSfs https://cdn.sdf.com/releases/download/install.sh | sh -s

Verify Installation

To verify that SDF is installed correctly, run sdf --help

SDF: A fast SQL compiler, local development framework, and in-memory analytical database

Usage: sdf [OPTIONS] <COMMAND>

Commands:   new      Create a new sdf workspace   clean    Remove artifacts that sdf has generated in the past   compile  Compile models   lint     Lint models   format   Format models   lsp1     Lsp wave1   run      Run models   build    Run model and tests, and if successful, publish, aka “write-audit-publish”   test     Test your models   stats    Statistics for your data   report   Report code quality   check    Check code quality   lineage  Display lineage for a given table and/or column   push     Push a local workspace to the SDF Service   system   System maintenance, install and update   auth     Authenticate CLI to services like SDF, AWS, OpenAI, etc   man      Display reference material, like the CLI, dialect specific functions, schemas for authoring and interchange   init     Initialize a workspace interactively   dbt      Initialize an sdf workspace from an existing dbt project   exec     Execute custom scripts   help     Print this message or the help of the given subcommand(s)

Options:       —log-level <LOG_LEVEL>  Set log level [possible values: trace, debug, debug-pretty, info, warn, error]       —log-file <LOG_FILE>    Creates or replaces the log file       —show-all-errors        Don’t suppress errors   -h, —help                   Print help   -V, —version                Print version

Use an editor like Visual Studio Code for better auto-complete and development hints.

YML Extension For VSCode

If you’re using VSCode, please download the official Red Hat YML extension which can provide intelisesnse support for SDF’s official YML schema. This extension will show configuration errors, options, and more.

The JSON schemas for SDF’s .sdf.yml extension are published on the schema store and in the sdf-cli repository.

SDF Versions and Updating

From time to time, it may be necessary to update the CLI to get the latest and greatest features, or to connect to SDF Cloud Services.

If you’ve already installed SDF, you can update to the latest version by running the command below. By default, this will update to the latest version within your current release channel (stable or preview).

sudo sdf system update

You can also update to a specific release channel: To view the available versions please see the github releases page.

sudo sdf system update --version x.x.x

You can check the currently installed version of your CLI with:

sdf --version

Release Tracks

SDF has two release tracks: stable and preview

stable (X.X.X-s) -> receives only explicit bug fixes, and is updated monthly

preview (X.X.X-p) -> receives new features multiple times per week

Installing a Release Channel Directly

You may install a preview version of SDF directly:

# To install a specific version of SDF
curl -LSfs https://cdn.sdf.com/releases/download/install.sh | sh -s - --version x.x.x

#To install the latest preview version
curl -LSfs https://cdn.sdf.com/releases/download/install.sh | sh -s - --preview

Switching Release Channels

Any SDF installation may switch tracks by updating into a different track, or version.

To switch channels while updating, use the --stable or --preview flags:

# Update to the latest stable version and switch to the stable channel
sudo sdf system update --stable

# Update to the latest preview version and switch to the preview channel
sudo sdf system update --preview

Preview versions offer early access to new features but may be less stable. Use for testing and feedback.

Uninstalling SDF

If you ever need to uninstall SDF, simply run

sudo sdf system uninstall

Info

SDF is currently supported on the following platforms. The executable is small, but heavily parallelized. As such, SDF’s resource consumption is dependent entirely on your data processing requirements.

PlatformsSupported
Linux Intel X86-64✔️
Linux Arm AARCH-64✔️
Apple Intel X86-64✔️
Apple Arm AARCh-64✔️
Windows Intel X86-64
Web-Assembly

For a complete list of releases, please see the official github releases page.

SDF is hardware optimized. For best performance, download the version that matches your chipset.