How to validate data without a Checkpoint

This guide demonstrates how to load an Expectation Suite and validate data without using a Checkpoint. This might be suitable for environments or workflows where a user does not want to or cannot create a Checkpoint, e.g. in a hosted environment.

Docs for Legacy Checkpoints (<=0.13.7)

Prerequisites: This how-to guide assumes you have already:

The following code mirrors the code provided in the validation_playground.ipynb notebooks in great_expectations/notebooks. First of all, we import Great Expectations, load our Data Context, and define variables for the Datasource we want to access:

import great_expectations as gx
context = gx.data_context.DataContext()

datasource_name = "my_datasource"

We then create a Batch using the above arguments. The batch_kwargs differ based on the type of data asset you want to connect to. The first example demonstrates the different possible batch_kwargs you could use to define your data for a SQLAlchemy Datasource:

# If you would like to validate an entire table or view in your database's default schema:
batch_kwargs = {'table': "YOUR_TABLE", 'datasource': datasource_name}

# If you would like to validate an entire table or view from a non-default schema in your database:
batch_kwargs = {'table': "YOUR_TABLE", "schema": "YOUR_SCHEMA", 'datasource': datasource_name}

# If you would like to validate the result set of a query:
batch_kwargs = {'query': 'SELECT YOUR_ROWS FROM YOUR_TABLE', 'datasource': datasource_name}

The following batch_kwargs can be used to create a batch for a Pandas or PySpark Datasource:

# If you would like to validate a file on a filesystem:
batch_kwargs = {'path': "YOUR_FILE_PATH", 'datasource': datasource_name}

# If you would like to validate in a PySpark or Pandas dataframe:
batch_kwargs = {'dataset': "YOUR_DATAFRAME", 'datasource': datasource_name}

Finally, we create the batch using those batch_kwargs and the name of the Expectation Suite we want to use, and run validation:

batch = context.get_batch(batch_kwargs, "my_expectation_suite")

results = context.run_validation_operator(
    run_id="my_run_id") # Make my_run_id a unique identifier, e.g. a timestamp

This runs validation and executes any ValidationActions configured for this ValidationOperator (e.g. saving the results to a ValidationResult Store).

Docs for Class-Based Checkpoints (>=0.13.8)


As part of the new modular expectations API in Great Expectations, Validation Operators have evolved into Class-Based Checkpoints. This means running a Validation without a Checkpoint is no longer supported in Great Expectations version 0.13.8 or later. Please read Checkpoints and Actions to learn more.