Soda SQL and Soda Spark are deprecated in favor of Soda Core and the Soda Checks Language. If you are new to Soda, start with Soda Core!
Data testing, monitoring, and profiling for SQL-accessible data.
){kind=icon}
✔ Install Soda SQL from the command-line
✔ Access comprehensive Soda SQL documentation
✔ Compatible with Snowflake, Amazon Redshift, BigQuery, and more
✔ Write tests in a YAML file
✔ Run programmatic scans to test data quality
table_name: breakdowns
metrics:
- row_count
- missing_count
- missing_percentage
...
# Validates that a table has rows
tests:
- row_count > 0
# Tests that numbers in the column are entered in a valid format as whole numbers
columns:
incident_number:
valid_format: number_whole
tests:
- invalid_percentage == 0
# Tests that no values in the column are missing
school_year:
tests:
- missing_count == 0
# Tests for duplicates in a column
bus_no:
tests:
- duplicate_count == 0
# Compares row count between datasets
sql_metric:
sql: |
SELECT COUNT(*) as other_row_count
FROM other_table
tests:
- row_count == other_row_countThanks goes to these wonderful people! (emoji key)
This project followed the all-contributors specification.
Soda SQL collects statistical usage and performance information via the Open Telemetry framework. This information helps the Soda Core developer team proactively track performance issues and understand how users interact with the tool.
The collected information is strictly limited to usage and performance and does not contain Personal Identifying Information (PII). It is used for internal purposes only. Soda keeps the data in its raw form for a maximum of five years. If some information needs to be kept for longer, it will only be kept in aggregated form.
Read more about the information Soda tracks, and learn how to opt-out of tracking by consulting the Anonymous usage statistics documentation.