dlt-runtime

0.24.1 · active · verified Fri Apr 17

dlt-runtime provides the `dlt` command-line interface (CLI) for the `dlt` data loading library. It allows users to initialize, run, deploy, and manage data pipelines built with `dlt`. The current version is 0.24.1, and it follows the frequent release cadence of the main `dlt` library, often with multiple minor releases per month.

Common errors

```
dlt: command not found
```
cause The `dlt-runtime` package, which installs the `dlt` CLI executable, was either not installed or not installed in an environment whose `bin` directory is on your system's PATH.

fix Ensure `dlt-runtime` is installed (`pip install dlt-runtime`). If it is, verify your shell's PATH includes the Python `Scripts` or `bin` directory where `dlt` is installed. For virtual environments, ensure the environment is activated.
```
ModuleNotFoundError: No module named 'dlt'
```
cause You installed `dlt-runtime` and are trying to `import dlt` but the `dlt` core library might be missing or corrupted, or you're trying to import from `dlt_runtime` directly.

fix While `dlt-runtime` depends on `dlt`, ensure the `dlt` package itself is correctly installed. If you are importing, always use `from dlt import ...`. If the error persists, try reinstalling `pip install dlt` (which `dlt-runtime` will pull anyway).
```
ERROR: dlt-runtime has requirement Python '>=3.9.2,<3.15', but you have Python 3.8.10 which is incompatible.
```
cause Your current Python environment does not meet the version requirements specified by the `dlt-runtime` package.

fix Upgrade your Python version to a compatible one (e.g., 3.9, 3.10, 3.11, 3.12, 3.13, 3.14 but not 3.15+) or switch to a Python environment that satisfies these requirements. Using `pyenv` or `conda` can help manage multiple Python versions.

Warnings

gotcha Common confusion: `dlt-runtime` provides the `dlt` CLI command, while the core Python library for building pipelines is the `dlt` package. While `dlt-runtime` depends on `dlt`, programmatic imports for pipeline development should always be `from dlt import ...`, not `from dlt_runtime import ...`.
Fix: Always use `from dlt import ...` for programmatic use of the `dlt` library. Use `dlt` command in your terminal for CLI operations.
breaking The `dlt` ecosystem (and thus the `dlt` CLI provided by `dlt-runtime`) is under active development. Frequent releases may introduce breaking changes, especially in major or minor versions, related to destination configurations, pipeline resource definitions, or CLI command syntax.
Fix: Refer to the official `dlt` documentation and release notes before upgrading, particularly for changes in `dlt` and related packages. Test your pipelines thoroughly after an upgrade.
gotcha Python version compatibility. `dlt-runtime` has specific Python version requirements (e.g., `>=3.9.2, <3.15` for version 0.24.1). Using an unsupported Python version will lead to installation or runtime errors.
Fix: Ensure your Python environment meets the requirements specified on the PyPI page for `dlt-runtime`. Use a tool like `pyenv` or `conda` to manage Python versions if needed.

Install

pip install dlt-runtime Install dlt-runtime

Imports

pipeline
wrong:
```
from dlt_runtime import pipeline
```
correct:
```
from dlt import pipeline
```
The `dlt-runtime` package provides the `dlt` CLI executable. For programmatic pipeline creation and data loading, you should `import dlt` from the `dlt` library itself (which `dlt-runtime` installs as a dependency). Do not try to import core `dlt` symbols directly from `dlt_runtime`.

Quickstart

This quickstart demonstrates how to use the `dlt` CLI (provided by `dlt-runtime`) to initialize, configure, and run a data loading pipeline. It uses `duckdb` as an example destination. Remember to adapt the pipeline script for your specific data sources.

# 1. Initialize a dlt project (e.g., using duckdb destination)
dlt init duckdb_pipeline duckdb

# 2. Navigate into the created project directory
cd duckdb_pipeline

# 3. Modify the pipeline script (e.g., duckdb_pipeline.py) if needed
#    For example, to load data from a source:
#    from dlt import pipeline, dlt_source
#    from my_source import my_data_source
#    
#    @dlt_source
#    def my_source():
#        yield [1, 2, 3] # Replace with actual data loading
#
#    @pipeline(pipeline_name='duckdb_pipeline', destination='duckdb')
#    def my_pipeline():
#        my_source() # Load data from the source

# 4. Run the pipeline
dlt run duckdb_pipeline.py

# 5. Check the status or deploy the pipeline
dlt status

view raw JSON →