asqlcell 0.3.1.0

Analytical sql cell for Jupyter

Analytical SQL Cell

For the last decade, the data ecosystem has mainly focused on the technologies to store and process big datasets — the bigger the better. Later on, Modern Data Stack emerged as a cloud-native suite of products used for data integration and data analytics by the more technology-forward companies. Data warehouses are now a default piece of the Modern Data Stack and Snowflake’s rapid rise has been the poster child of this trend.

But in real life, most analytical workloads aren’t massive. Users prefer easy and fast answers to their questions instead of waiting for the cloud to spin. Instead of a distributed database in the cloud, most analyses can be handled with an optimized engine on our laptop and the cloud can be leveraged when needed.

To fill the gap, we built Analytical SQL Cell, a free and open-source Jupyter Widget that offers a Personal Data Lake experience for Jupyter Notebook users.

Analytical SQL has the following features:

• Pain-free setup (pip install asqlcell and you are in).
• Unleash the power of SQL for analytical workloads.
• Lightning speed for data loading.
• Built-in visualization.
• Low footprint for analytical processing.
• Interleave Python and SQL programming.
• Free and open source.

Installation

You can install using pip:

pip install asqlcell

Quick Start

Here's a quick example to get you started with Analytical SQL Cell in Jupyter:

%%sql

SELECT
    continent,
    AVG(life_exp) AS health,
    AVG(gdp_cap) AS wealth,
    SUM(population) AS population
FROM 'gapminder.csv.gz'
WHERE year=2007
GROUP BY continent

In the above sample, %%sql is a cell magic indicating that the cell block will be executed as a SQL statement. File are used as data source for querying with SQL.

Data Load

You can query from Pandas DataFrame, CSV files, compressed (e.g. compressed with gzip) CSV files, as well as Parquet files.

SQL

DuckDB is the default engine of Analytical SQL Cell. Please find more details at the SQL Introduction of DuckDB.

Result Table

With the SQL query being executed in an Analytical SQL Cell, the result set is presented as a table.

Chart Builder

You can further visualize the result set by clicking the Chart tab.

Here are the chart types supported:

• Column chart
• Bar chart
• Line chart
• Area chart
• Scatter chart
• Pie chart
• Combo chart
• Funnel chart
• Sunburst chart

Tutorial

• Introduction
• World Development in Numbers
• Visualise Netflix Data With Analytical SQL
• Learn SQL with Jupyter Lab
• Business Questions on Chinook Music Store
• Use the Right Chart for Data Visualization

Development

This widget is developed with conda to ensure a consistent developer experience. The project is developed in both Mac OS and Windows Subsystem for Linux.

Please run the following commands to create a conda environment:

conda create -n asqlcell -c conda-forge nodejs=18.15 python=3.8 jupyterlab=3.6 jupyter_packaging=0.12
conda activate asqlcell

Python

Install the python. This will also build the TypeScript package.

pip install -e ".[test, examples, docs]"

Jupyter

When developing your extensions, you need to manually enable your extensions with the notebook / lab frontend.

For lab, this is done by the command:

jupyter labextension develop --overwrite .
jlpm run build

Typescript

You must start watching the change of the widget:

jlpm run watch

Then in another terminal you can run Jupyter Lab (without launching a browser):

jupyter lab --no-browser

Now you can open Google Chrome and navigate to http://localhost:8888 to play with the widget. After a change wait for the build to finish and then refresh your browser and the changes should take effect.

If you make a change to the python code then you will need to restart the notebook kernel to have it take effect.