Why does this result page does not count failed expectation test? But if I send true values it shows it in the resulting page.

[almatkai]

import great_expectations as gx

context = gx.get_context()

validator = context.sources.pandas_default.read_csv(
    "data.csv"
)

res1 = validator.expect_column_values_to_not_be_null("pickup_datetime")
# This one below with Error(true => min=1, max=6)
res2 = validator.expect_column_values_to_be_between("passenger_count", min_value=1, max_value=5)
validator.save_expectation_suite()

checkpoint = context.add_or_update_checkpoint(
    name="my_quickstart_checkpoint",
    validator=validator,
)

checkpoint_result = checkpoint.run()
context.view_validation_result(checkpoint_result)

[michalc]

(I have literally 1 day of experience with Great Expectations)

This issue just surprised us - we were trying to make a failing expectation to make sure the system worked, and it would just disappear.

To fix this, we changed:

validator.save_expectation_suite()

to include discard_failed_expectations=False

validator.save_expectation_suite(discard_failed_expectations=False)

It's in the docs at: https://docs.greatexpectations.io/docs/reference/api/validator/validator/Validator_class#great_expectations.validator.validator.Validator.save_expectation_suite

I'm starting to suspect something: the documentation assumes you have a "correct" dataset that you use to interactively develop expectations with. It's not really our use case though - we might have "wrong data", or even in fact, we might not even have any data at all.

(Right now we're kinda wrestling with trying to work out how to use Great Expectations for our case, if at all - the assumptions/workflows don't quite seem aligned with our world... Do we adjust our workflows, or can we figure out how to use Great Expectations within them?)