Skip to main content

Asserts

re_data comes with asserts library, created to test metrics computed. You can use those as tests in schema.yml of every model monitored by re_data.

re_data asserts can be applied both as table and column level tests as show in the example.

schema.yml
version: 2

models:
  - name: buy_events
    tests:
      - re_data.assert_in_range:
          metric: row_count
          min_value: 0
          max_value: 10

    columns:
      - name: value1
        tests:
          - re_data.assert_true:
              metric: nulls_percent
              expression: value = 0

      - name: value2
        tests:
          - re_data.assert_greater:
              metric: min
              value: 200
              condition: time_window_start >= '2020-05-02'

info

Asserts have a set of standard params with the same meaning in all tests:

  • metric: name of the metric you are testing
  • expression: an expression which re_data will evaluate to be true or false. Use value to indicate the value of a metric computed. Example expression: value > 0 asserts that metric is greater than 0
  • value - a number to which you would like to compare metric. Example: value: 5 in the assert_greater test would check if all metric values are larger than 5
  • min_value, max_value - similarly to value re_data will compare metric values to values passed here
  • condition: (optional) time filter for the metric, if you would like to tests only metric from specific time range. User time_window_start or time_window_end to compare against time. Example: time_window_start > '2020-05-02' will check only metrics computed later than on '2020-05-02'

assert_true​

Accepted params:

  • metric
  • expression
  • condition

Assert that given expression is true for the metric computed.

- re_data.assert_true:
    metric: nulls_percent
    expression: value = 0
    condition: time_window_start >= '2020-05-02' # (optinal)

assert_false​

Accepted params:

  • metric
  • expression
  • condition

Assert that given expression is false for the metric computed.

- re_data.assert_false:
    metric: freshness
    expression: value is null

assert_in_range​

Accepted params:

  • metric
  • min_value
  • max_value
  • condition

Assert that metric value is between min_value and max_value (inclusive)

- re_data.assert_in_range:
    metric: row_count
    min_value: 0
    max_value: 10

assert_equal​

Accepted params:

  • metric
  • value
  • condition

Assert that metric value is equal to specified value

- re_data.assert_equal:
    metric: row_count
    value: 1

assert_greater​

Accepted params:

  • metric
  • value
  • condition

Assert that metric value is greater to specified value

- re_data.assert_greater:
    metric: min
    value: 200

assert_greater_equal​

Accepted params:

  • metric
  • value
  • condition

Assert that metric value is greater or equal specified value

- re_data.assert_greater_equal:
    metric: my_distinct_table_rows
    value: 10

assert_less​

Accepted params:

  • metric
  • value
  • condition

Assert that metric value is smaller specified value

- re_data.assert_less:
    metric: min
    value: 100

assert_less_equal​

Accepted params:

  • metric
  • value
  • condition

Assert that metric value is smaller or equal to specified value

- re_data.assert_less_equal:
    metric: min
    value: 107

If you would like us to add some other tests, information about tests history, etc., join Slack! 😊 and let us know! 😊