CSV

CSVs are a simple file format that stores data in a human-readable, plain text format. They are widely used for data storage and transfer. Evidence supports connecting to CSV files as a data source, allowing you to query them using SQL.

Add a New CSV Data Source

1. Start the Evidence dev server: npm run dev or Start Evidence using the VSCode extension.
2. Navigate to the settings page, also accessible via the ... menu in the top right.
3. Click the New Source button.
4. Select CSV as the data source type.
5. Enter the configuration options for the connection.
6. Click the Test button to confirm the connection is successful.
7. Click the Save button to save the connection.
8. Your configuration options are saved in two files:
   • /sources/[source_name]/connections.yaml: Non-sensitive values, source controlled.
   • /sources/[source_name]/connection.options.yaml: Sensitive values, not source controlled, base-64 encoded.

Then copy any CSV files you want to query into sources/[your_csv_source_name]/. Your source names and csv files can only contain letters, numbers and underscores eg /my_source/my_csv_2024.csv

How to Query a CSV File

Evidence looks for files with the .csv extension stored in a sources/[your_csv_source_name]/ folder in the root of your Evidence project. You can query them using this syntax:

select * from your_csv_source_name.csv_file_name

Do not include the .csv extension in the file name when querying.

Configuration

You can add DuckDB source options that are passed in as arguments to the read_csv() function.

Ensure there are no spaces in your source options you pass, and to use double quotes when passing strings