Great Expectations (GX)

Visit Website

Introduction

Great Expectations, now called GX Core for the open-source library, says it helps teams ‘always know what to expect from your data’. Its README describes powerful technical tools that start with Expectations: expressive and extensible unit tests for your data. Expectations give teams a common language to express data quality tests in an intuitive way, and documentation can be generated automatically for each set of validation results, which helps preserve an organisation’s knowledge about its data.

An Expectation is a declarative assertion such as ‘values in this column are not null’, ‘values are between 0 and 100’, ‘this column matches a regex’, or ‘the table has between 1,000 and 2,000 rows’. Expectations are grouped into suites and run against batches of data from sources such as pandas and Spark DataFrames and SQL databases through a Data Context. Validation definitions and checkpoints run suites and trigger actions, for example updating documentation or sending notifications, and results are rendered as Data Docs, human-readable HTML reports.

For software testers, GX extends testing into the data layer. Many defects in data-heavy systems are not code bugs but data bugs: a source system sends nulls, a join duplicates rows, or a unit changes. Running Expectations in pipelines catches these before reports or downstream services consume bad data, and the same checks can verify test datasets and migrations.

GX Core supports Python 3.10 to 3.14, with experimental support for newer versions behind an environment variable. It has about 11,900 GitHub stars, released 1.24.1 in October 2026, and had commits on 11 November 2026. GX Cloud, the commercial hosted service, offers a free Developer option and Team and Enterprise plans, according to the pricing page.

Support from the project is prioritised through GitHub issues and pull requests, then the Discourse forum, then the Slack community.

In the quality loop, Great Expectations sits at integration and release stages for data pipelines, and in production as ongoing data monitoring.

The source code is at fivetran/great_expectations on GitHub under the Apache-2.0 licence, with about 11,900 stars; the most recent commit was on 11 November 2026.

Choosing between GX and lighter tools depends on scale and audience. Pandera suits Python teams validating dataframes inside code, Soda suits teams who prefer YAML contracts for warehouses, and Deequ suits Spark-heavy platforms. GX sits in the middle: it is code-first like Pandera but adds suites, checkpoints and documentation that make data quality visible to people who do not read the code. A pragmatic approach is to begin with the checks that would have caught your last few data incidents, run them on every pipeline execution, and grow the suite as new failure modes appear. Treat failing Expectations as triage signals, deciding for each one whether the data or the Expectation needs to change.

Features

1. Expectations: Declarative tests for data.
2. Expectation suites: Group related checks.
3. Many data sources: pandas, Spark and SQL databases.
4. Checkpoints and actions: Run validations and trigger notifications.
5. Data Docs: Auto-generated HTML documentation.
6. Extensible: Write custom Expectations.
7. Expectation Gallery: Catalogue of available checks.
8. Pipeline integration: Run from orchestrators and CI.
9. GX Cloud: Hosted option with a free Developer tier.
10. Active community: Forum and Slack.

Pros

1. Readable data tests: Shared language for teams.
2. Documentation built in: Data Docs from results.
3. Free to use: Open source under the Apache-2.0 licence, so the core tool costs nothing to use.
4. Actively maintained: The repository had commits as recent as 11 November 2026.
5. Community: About 11,900 GitHub stars.

Cons

1. Learning curve: Concepts like Data Context and checkpoints take time.
2. API changes: 1.x differs from older 0.x tutorials.
3. Python only: Other stacks need wrappers.

Usage and Links (Example)

pip install great_expectations

import great_expectations as gx
import pandas as pd

context = gx.get_context()
df = pd.read_csv("orders.csv")
batch = context.data_sources.pandas_default.read_dataframe(df)

result = batch.validate(gx.expectations.ExpectColumnValuesToNotBeNull(column="order_id"))
print(result.success)

Practical tips:

1. Start with a few high-value checks on critical tables, such as keys, nulls and value ranges.
2. Run validations in CI for transformation code and in production pipelines for incoming data.
3. Use Data Docs to share results with analysts and product owners.
4. Make sure tutorials you follow are for GX 1.x.

Resources:

• Great Expectations (GX) website
• GitHub repository

Pricing

Great Expectations (GX) is free and open source under the Apache-2.0 licence.

Recommended for?

1. Data engineers and QA teams testing data pipelines.
2. Analytics teams wanting documented data quality.
3. Not a fit for testing application UI or APIs.

Conclusion

Great Expectations is the best-known open-source framework for data quality testing, with a clear model of Expectations and built-in documentation.

Expect some learning curve, and keep to 1.x documentation, but the payoff is reliable, well-documented data checks.

On CheckOps, Great Expectations (GX) is filed under the Integration testing, Release & CI gates and Production monitoring & observability stages. Related tools listed here include Soda Core, Deequ, Pandera, pytest, Hypothesis and Giskard. Before choosing, compare them on licence, supported languages and platforms, hosting model, and how well each fits your existing CI pipeline. Facts on this page were checked against the official website, documentation and repository on 11 November 2026; statements the vendor makes about its own product are marked as vendor claims.