Shelect: SQL for local files

Shelect lets you run SQL on data in your local filesystem.

Use it for easy, local ad-hoc analysis of CSV/TSV/JSON files.

It supports all functions and SQL syntax supported by SQLite.

How to use it

$ shelect 'SELECT * FROM "./examples/people.csv"'
id | name
---+-------
1  | Alice
2  | Bob
3  | Carlos
4  | Dani
$ shelect 'SELECT * FROM "./examples/values.json"'
id | value
---+------
1  | 10
2  | 20
3  | 30
3  | 40
$ shelect 'SELECT * FROM "./examples/values.json" JOIN "./examples/people.csv" USING (id)'
id | value | name
---+-------+-------
1  | 10    | Alice
2  | 20    | Bob
3  | 30    | Carlos
3  | 40    | Carlos
$ shelect 'SELECT name, sum(value) FROM "./examples/people.csv" LEFT JOIN "./examples/values.json" USING (id) GROUP BY 1'
name   | SUM(value)
-------+-----------
Alice  | 10
Bob    | 20
Carlos | 70
Dani   | NULL

It also can function as a repl when invoked without a query.

How it works

An in-memory SQLite database is created. Before executing each query, the query is parsed and all tables accessed are first imported from the filesystem into a table with that name.

So if you run SELECT * from "./my-file.csv" a table named "./my-file.csv" will be created by opening ./my-file.csv and attempting to read it as JSON, CSV, or TSV.

JSON files must be arrays of objects, where each object is a key-value mapping where keys correspond to column names.

CSV and TSV files must have a single row with column names followed by data rows.

All columns are created as TEXT data types. Please refer to the SQLite documentation for supported SQL functions and functionality.

How to install

shelect is on pypi: https://pypi.org/project/shelect/

Via pipx:

$ pipx install shelect

Via pip:

$ python3 -m pip install shelect