Evaluation Parameter Store (Metric Store)
Overview
Definition
An Evaluation Parameter Store is a connector to store and retrieve information about parameters used during Validation of an Expectation which reference simple expressions or previously generated metrics.
Features and promises
Evaluation Parameter Stores provide a method of storing and retrieving the Metrics associated with Evaluation Parameters without also storing the additional information that accompanies Validation Results.
Relationship to other objects
Evaluation Parameter Stores are an alternative way to
store Metrics without the accompanying Validation
Results. The process of storing information in an
Evaluation Parameter Store is also typically executed
by an Action in a Checkpoint's
action_list
. Metrics stored in Evaluation
Parameter Stores are also available as Evaluation
Parameters when defining Expectations.
Use cases
Setup |
When you initialize your Data Context for the first
time, a configuration for an in-memory Evaluation
Parameter Store will automatically be added to
great_expectations.yml
. You may change
this configuration to work with different environments
in the same fashion as you would configure any other
store, by adding a store_backend
key to
the Evaluation Parameter Store's entry.
Create Expectations |
When some Expectations are created they require you to
define an Evaluation Parameter. One of the ways this
can be done is to define a
'$PARAMETER': URN
pair in the
Expectation's value
dictionary. When
Great Expectations encounters a $PARAMETER flag during
Validation, it will replace the URN
with
a value retrieved from an Evaluation Parameter Store
or Metric Store.
Validate Data |
If an Evaluation Parameter is used in an Expectation that is being Validated, and the Evaluation Parameter is configured to reference a value in your Evaluation Parameter Store, the Evaluation Parameter Store will be used behind the scenes to retrieve that value. You will not need to directly interact with it.
If a Checkpoint is run,and it contains the
StoreEvaluationParametersAction
Action in
its action_list
, then your Evaluation
Parameter Store will be used to store the Evaluation
Parameters that were generated by the Checkpoint as
well as their values.
Features
Referencing prior results
The Evaluation Parameter Store permits you to store the Metrics generated by Evaluation Parameters when Expectations containing them are Validated. It also allows you to retrieve these Evaluation Parameters and values at a later point. This permits you to create Expectations that rely on comparing the data of your current Batch with the metrics generated from a previous Batch of data.
Convenience of use
Reading from and writing to the Evaluation Parameter
store is handled behind the scenes. All you need to do
is provide the correct URN path to the parameter you
want to retrieve, or include the
StoreEvaluationParametersAction
Action in
the action_list
of any Checkpoint that
generates Metrics from Evaluation Parameters that you
want to store.
API basics
How to access
The Evaluation Parameter Store uses an URN schema for
identifying dependencies between Expectation Suites. A
valid URN must begin with the string:
urn:great_expectations
. Valid URNs for
Evaluation Parameter Stores must have one of the
following structures to be recognized by your Data
Context:
urn:great_expectations:validations:<expectation_suite_name>:<metric_name>
urn:great_expectations:validations:<expectation_suite_name>:<metric_name>:<metric_kwargs_id>