Skip to main content

How to configure an InferredAssetDataConnector

This guide demonstrates how to configure an InferredAssetDataConnector, and provides several examples you can use for configuration.

Prerequisites: This how-to guide assumes you have:

Great Expectations provides two types of DataConnector classes for connecting to Data Assets stored as file-system-like data (this includes files on disk, but also S3 object stores, etc) as well as relational database data:

  • A ConfiguredAssetDataConnector allows you to specify that you have multiple Data Assets in a Datasource, but also requires an explicit listing of each Data Asset you want to connect to. This allows more fine-tuning, but also requires more setup.
  • An InferredAssetDataConnector infers data_asset_name by using a regex that takes advantage of patterns that exist in the filename or folder structure.

InferredAssetDataConnector has fewer options, so it's simpler to set up. It’s a good choice if you want to connect to a single Data Asset, or several Data Assets that all share the same naming convention.

If you're not sure which one to use, please check out How to choose which DataConnector to use.

Steps#

1. Instantiate your project's DataContext#

Import these necessary packages and modules:

from ruamel import yaml
import great_expectations as gefrom great_expectations.core.batch import BatchRequest

2. Set up a Datasource#

All the examples below assume you’re testing configurations using something like:

datasource_yaml = """name: taxi_datasourceclass_name: Datasourceexecution_engine:  class_name: PandasExecutionEnginedata_connectors:  <DATA CONNECTOR NAME GOES HERE>:    <DATACONNECTOR CONFIGURATION GOES HERE>"""context.test_yaml_config(yaml_config=datasource_config)

If you’re not familiar with the test_yaml_config method, please check out: How to configure Data Context components using test_yaml_config

3. Add an InferredAssetDataConnector to a Datasource configuration#

InferredAssetDataConnectors like InferredAssetFilesystemDataConnector and InferredAssetS3DataConnector require a default_regex parameter, with a configured regex pattern and capture group_names.

Imagine you have the following files in my_directory/:

<MY DIRECTORY>/yellow_tripdata_2019-01.csv<MY DIRECTORY>/yellow_tripdata_2019-02.csv<MY DIRECTORY>/yellow_tripdata_2019-03.csv

We can imagine two approaches to loading the data into GE.

The simplest approach would be to consider each file to be its own Data Asset. In that case, the configuration would look like the following:

datasource_yaml = r"""name: taxi_datasourceclass_name: Datasourcemodule_name: great_expectations.datasourceexecution_engine:  module_name: great_expectations.execution_engine  class_name: PandasExecutionEnginedata_connectors:  default_inferred_data_connector_name:    class_name: InferredAssetFilesystemDataConnector    base_directory: <MY DIRECTORY>/    default_regex:      group_names:        - data_asset_name      pattern: (.*)\.csv"""

Notice that the default_regex is configured to have one capture group ((.*)) which captures the entire filename. That capture group is assigned to data_asset_name under group_names. For InferredAssetDataConnectors data_asset_name is a required group_name, and it's associated capture group is the way each data_asset_name is inferred. Running test_yaml_config() would result in 3 Data Assets : yellow_tripdata_2019-01, yellow_tripdata_2019-02 and yellow_tripdata_2019-03.

However, a closer look at the filenames reveals a pattern that is common to the 3 files. Each have yellow_tripdata_ in the name, and have date information afterwards. These are the types of patterns that InferredAssetDataConnectors allow you to take advantage of.

We could treat yellow_tripdata_*.csv files as batches within the yellow_tripdata Data Asset with a more specific regex pattern and adding group_names for year and month.

Note: We have chosen to be more specific in the capture groups for the year and month by specifying the integer value (using \d) and the number of digits, but a simpler capture group like (.*) would also work. For more information about capture groups, refer to the Python documentation on regular expressions.

datasource_yaml = r"""name: taxi_datasourceclass_name: Datasourcemodule_name: great_expectations.datasourceexecution_engine:  module_name: great_expectations.execution_engine  class_name: PandasExecutionEnginedata_connectors:  default_inferred_data_connector_name:    class_name: InferredAssetFilesystemDataConnector    base_directory: <MY DIRECTORY>/    default_regex:      group_names:        - data_asset_name        - year        - month      pattern: (.*)_(\d{4})-(\d{2})\.csv"""

Running test_yaml_config() would result in 1 Data Asset yellow_tripdata with 3 associated data_references: yellow_tripdata_2019-01.csv, yellow_tripdata_2019-02.csv and yellow_tripdata_2019-03.csv, seen also in Example 1 below.

A corresponding configuration for InferredAssetS3DataConnector would look similar but would require bucket and prefix values instead of base_directory.

datasource_yaml = r"""name: taxi_datasourceclass_name: Datasourcemodule_name: great_expectations.datasourceexecution_engine:  module_name: great_expectations.execution_engine  class_name: PandasExecutionEnginedata_connectors:  default_inferred_data_connector_name:    class_name: InferredAssetS3DataConnector    bucket: <MY S3 BUCKET>/    prefix: <MY S3 BUCKET PREFIX>/    default_regex:      group_names:        - prefix        - data_asset_name        - year        - month      pattern: (.*)/(.*)_sample_(\d{4})-(\d{2})\.csv"""

The following examples will show scenarios that InferredAssetDataConnectors can help you analyze, using InferredAssetFilesystemDataConnector.

Example 1: Basic configuration for a single Data Asset#

Continuing the example above, imagine you have the following files in the directory <MY DIRECTORY>:

<MY DIRECTORY>/yellow_tripdata_2019-01.csv<MY DIRECTORY>/yellow_tripdata_2019-02.csv<MY DIRECTORY>/yellow_tripdata_2019-03.csv

Then this configuration:

datasource_yaml = r"""name: taxi_datasourceclass_name: Datasourcemodule_name: great_expectations.datasourceexecution_engine:  module_name: great_expectations.execution_engine  class_name: PandasExecutionEnginedata_connectors:  default_inferred_data_connector_name:    class_name: InferredAssetFilesystemDataConnector    base_directory: <MY DIRECTORY>/    default_regex:      group_names:        - data_asset_name        - year        - month      pattern: (.*)_(\d{4})-(\d{2})\.csv"""

will make available yelow_tripdata as a single Data Asset with the following data_references:

Available data_asset_names (1 of 1):    yellow_tripdata (3 of 3): ['yellow_tripdata_2019-01.csv', 'yellow_tripdata_2019-02.csv', 'yellow_tripdata_2019-03.csv']
Unmatched data_references (0 of 0):[]

Once configured, you can get Validators from the Data Context as follows:

batch_request = BatchRequest(    datasource_name="taxi_datasource",    data_connector_name="default_inferred_data_connector_name",    data_asset_name="yellow_tripdata",)
validator = context.get_validator(    batch_request=batch_request,    create_expectation_suite_with_name="<MY EXPECTATION SUITE NAME>",)

Since this BatchRequest does not specify which data_reference to load, the ActiveBatch for the validator will be the last data_reference that was loaded. In this case, yellow_tripdata_2019-03.csv is what is being used by validator. We can verfiy this with:

print(validator.active_batch_definition)

which prints:

{  "datasource_name": "taxi_datasource",  "data_connector_name": "default_inferred_data_connector_name",  "data_asset_name": "yellow_tripdata",  "batch_identifiers": {    "year": "2019",    "month": "03"  }}

Notice that the batch_identifiers for this batch_definition specify "year": "2019", "month": "03". The parameter batch_identifiers can be used in our BatchRequest to return the data_reference CSV of our choosing using the group_names defined in our DataConnector:

batch_request = BatchRequest(    datasource_name="taxi_datasource",    data_connector_name="default_inferred_data_connector_name",    data_asset_name="yellow_tripdata",    data_connector_query={"batch_filter_parameters": {"year": "2019", "month": "02"}},)
validator = context.get_validator(    batch_request=batch_request,    expectation_suite_name="<MY EXPECTATION SUITE NAME>",)
print(validator.active_batch_definition)

which prints:

{  "datasource_name": "taxi_datasource",  "data_connector_name": "default_inferred_data_connector_name",  "data_asset_name": "yellow_tripdata",  "batch_identifiers": {    "year": "2019",    "month": "02"  }}

This ability to access specific Batches using batch_identifiers is very useful when validating Data Assets that span multiple files. For more information on batches and batch_identifiers, please refer to the Core Concepts document.

Example 2: Basic configuration with more than one Data Asset#

Here’s a similar example, but this time two Data Assets are mixed together in one folder.

Note: For an equivalent configuration using ConfiguredAssetFilesSystemDataconnector, please see Example 2 in How to configure a ConfiguredAssetDataConnector.

<MY DIRECTORY>/yellow_tripdata_2019-01.csv<MY DIRECTORY>/green_tripdata_2019-01.csv<MY DIRECTORY>/yellow_tripdata_2019-02.csv<MY DIRECTORY>/green_tripdata_2019-02.csv<MY DIRECTORY>/yellow_tripdata_2019-03.csv<MY DIRECTORY>/green_tripdata_2019-03.csv

The same configuration as Example 1:

datasource_yaml = r"""name: taxi_datasourceclass_name: Datasourcemodule_name: great_expectations.datasourceexecution_engine:  module_name: great_expectations.execution_engine  class_name: PandasExecutionEnginedata_connectors:  default_inferred_data_connector_name:    class_name: InferredAssetFilesystemDataConnector    base_directory: <MY DIRECTORY>/    default_regex:      group_names:        - data_asset_name        - year        - month      pattern: (.*)_(\d{4})-(\d{2})\.csv"""

will now make yellow_tripdata and green_tripdata both available as Data Assets, with the following data_references:

Available data_asset_names (2 of 2):    green_tripdata (3 of 3): ['green_tripdata_2019-01.csv', 'green_tripdata_2019-02.csv', 'green_tripdata_2019-03.csv']    yellow_tripdata (3 of 3): ['yellow_tripdata_2019-01.csv', 'yellow_tripdata_2019-02.csv', 'yellow_tripdata_2019-03.csv']
Unmatched data_references (0 of 0): []

Example 3: Nested directory structure with the data_asset_name on the inside#

Here’s a similar example, with a nested directory structure:

<MY DIRECTORY>/2018/10/yellow_tripdata.csv<MY DIRECTORY>/2018/10/green_tripdata.csv<MY DIRECTORY>/2018/11/yellow_tripdata.csv<MY DIRECTORY>/2018/11/green_tripdata.csv<MY DIRECTORY>/2018/12/yellow_tripdata.csv<MY DIRECTORY>/2018/12/green_tripdata.csv<MY DIRECTORY>/2019/01/yellow_tripdata.csv<MY DIRECTORY>/2019/01/green_tripdata.csv<MY DIRECTORY>/2019/02/yellow_tripdata.csv<MY DIRECTORY>/2019/02/green_tripdata.csv<MY DIRECTORY>/2019/03/yellow_tripdata.csv<MY DIRECTORY>/2019/03/green_tripdata.csv

Then this configuration:

datasource_yaml = r"""name: taxi_datasourceclass_name: Datasourcemodule_name: great_expectations.datasourceexecution_engine:  module_name: great_expectations.execution_engine  class_name: PandasExecutionEnginedata_connectors:  default_inferred_data_connector_name:    class_name: InferredAssetFilesystemDataConnector    base_directory: <MY DIRECTORY>/    glob_directive: "*/*/*.csv"    default_regex:      group_names:        - year        - month        - data_asset_name      pattern: (\d{4})/(\d{2})/(.*)\.csv"""

will now make yellow_tripdata and green_tripdata both available as Data Assets, with the following data_references:

Available data_asset_names (2 of 2):    green_tripdata (3 of 6): ['2018/10/green_tripdata.csv', '2018/11/green_tripdata.csv', '2018/12/green_tripdata.csv']    yellow_tripdata (3 of 6): ['2018/10/yellow_tripdata.csv', '2018/11/yellow_tripdata.csv', '2018/12/yellow_tripdata.csv']
Unmatched data_references (0 of 0):[]

The glob_directive is provided to give the DataConnector information about the directory structure to expect for each Data Asset. The default glob_directive for the InferredAssetFileSystemDataConnector is "*" and therefore must be overridden when your data_references exist in subdirectories.

Example 4: Nested directory structure with the data_asset_name on the outside#

In the following example, files are placed in a folder structure with the data_asset_name defined by the folder name (yellow_tripdata or green_tripdata)

<MY DIRECTORY>/yellow_tripdata/yellow_tripdata_2019-01.csv<MY DIRECTORY>/yellow_tripdata/yellow_tripdata_2019-02.csv<MY DIRECTORY>/yellow_tripdata/yellow_tripdata_2019-03.csv<MY DIRECTORY>/green_tripdata/2019-01.csv<MY DIRECTORY>/green_tripdata/2019-02.csv<MY DIRECTORY>/green_tripdata/2019-03.csv

Then this configuration:

datasource_yaml = r"""name: taxi_datasourceclass_name: Datasourcemodule_name: great_expectations.datasourceexecution_engine:  module_name: great_expectations.execution_engine  class_name: PandasExecutionEnginedata_connectors:  default_inferred_data_connector_name:    class_name: InferredAssetFilesystemDataConnector    base_directory: <MY DIRECTORY>/    glob_directive: "*/*.csv"    default_regex:      group_names:        - data_asset_name        - file_name_root        - year        - month      pattern: (.*)/(.*)(\d{4})-(\d{2})\.csv"""

will now make yellow_tripdata and green_tripdata into Data Assets, with each containing 3 data_references

Available data_asset_names (2 of 2):    green_tripdata (3 of 3): ['green_tripdata/2019-01.csv', 'green_tripdata/2019-02.csv', 'green_tripdata/2019-03.csv']    yellow_tripdata (3 of 3): ['yellow_tripdata/yellow_tripdata_2019-01.csv', 'yellow_tripdata/yellow_tripdata_2019-02.csv', 'yellow_tripdata/yellow_tripdata_2019-03.csv']
Unmatched data_references (0 of 0):[]

Example 5: Redundant information in the naming convention#

In the following example, files are placed in a folder structure with the data_asset_name defined by the folder name (yellow_tripdata or green_tripdata), but then the term yellow_tripdata is repeated in some filenames.

<MY DIRECTORY>/yellow_tripdata/yellow_tripdata_2019-01.csv<MY DIRECTORY>/yellow_tripdata/yellow_tripdata_2019-02.csv<MY DIRECTORY>/yellow_tripdata/yellow_tripdata_2019-03.csv<MY DIRECTORY>/green_tripdata/2019-01.csv<MY DIRECTORY>/green_tripdata/2019-02.csv<MY DIRECTORY>/green_tripdata/2019-03.csv

Then this configuration:

datasource_yaml = r"""name: taxi_datasourceclass_name: Datasourcemodule_name: great_expectations.datasourceexecution_engine:  module_name: great_expectations.execution_engine  class_name: PandasExecutionEnginedata_connectors:  default_inferred_data_connector_name:    class_name: InferredAssetFilesystemDataConnector    base_directory: <MY DIRECTORY>/    glob_directive: "*/*.csv"    default_regex:      group_names:        - data_asset_name        - year        - month      pattern: (.*)/.*(\d{4})-(\d{2})\.csv"""

will not display the redundant information:

Available data_asset_names (2 of 2):    green_tripdata (3 of 3): ['green_tripdata/*2019-01.csv', 'green_tripdata/*2019-02.csv', 'green_tripdata/*2019-03.csv']    yellow_tripdata (3 of 3): ['yellow_tripdata/*2019-01.csv', 'yellow_tripdata/*2019-02.csv', 'yellow_tripdata/*2019-03.csv']
Unmatched data_references (0 of 0):[]

Additional Notes#

To view the full script used in this page, see it on GitHub: