Create an Expectation Suite

[jwestw]

An Expectation Suite in Great Expectations is a collection of Expectations that apply to a specific dataset.

You can create a new suite programmatically. The name from your TOML's [data_asset] table can be used as the Expectation Suite name.

[james-westwood]

The task is to build a function or method that constructs a Great Expectation expectation suite from the inputs which are dictionaries from the generator functions.

Each of the Expectation chooser functions will return a dictionary something like this.

{
      "meta": {},
      "expectation_type": "example_expectation_name",
      "kwargs": {
        "column": "example_column_name"
      }

You can add an expectation to the suite using:

suite.expect_column_values_to_not_be_null(column)

This will add the "expect_column_values_to_not_be_null" expectation to the Great Expectations suite object. It modifies the suite in place. However, this change is only in memory until you explicitly save the suite using

context.save_expectation_suite(suite, "my_expectation_suite")

The saving action persists the expectation to a file, typically in JSON format, making it part of your saved Expectation Suite.

Ultimately an expectation suite will look like this.

{
  "expectation_suite_name": "example_survey_results_expectations",
  "ge_cloud_id": null,
  "expectations": [
    {
      "meta": {},
      "expectation_type": "expect_table_columns_to_match_ordered_list",
      "kwargs": {
        "column_list": ["reference", "period", "survey", "date_col_example_1", "date_col_example_2", "integer_col_example"]
      }
    },
    {
      "meta": {},
      "expectation_type": "expect_column_values_to_not_be_null",
      "kwargs": {
        "column": "reference"
      }
    },
    {
      "meta": {},
      "expectation_type": "expect_column_value_lengths_to_be_at_least",
      "kwargs": {
        "column": "reference",
        "min_value": 1
      }
    },
     {
      "meta": {},
      "expectation_type": "expect_column_values_to_match_regex",
      "kwargs": {
        "column": "reference",
        "regex": "^\\d{6}$"
      }
    },
    {
      "meta": {},
      "expectation_type": "expect_column_values_to_be_unique",
      "kwargs": {
        "column": "reference"
      }
    }
  ],
  "data_asset_type": "Dataset",
  "meta": {}
}