Skip to content

FeatureGroup#

[source]

FeatureGroup#

hsfs.feature_group.FeatureGroup(
    name,
    version,
    featurestore_id,
    description="",
    partition_key=None,
    primary_key=None,
    hudi_precombine_key=None,
    featurestore_name=None,
    created=None,
    creator=None,
    id=None,
    features=None,
    location=None,
    online_enabled=False,
    time_travel_format=None,
    statistics_config=None,
    validation_type="NONE",
    expectations=None,
    online_topic_name=None,
    event_time=None,
)

Creation#

[source]

create_feature_group#

FeatureStore.create_feature_group(
    name,
    version=None,
    description="",
    online_enabled=False,
    time_travel_format="HUDI",
    partition_key=[],
    primary_key=[],
    hudi_precombine_key=None,
    features=[],
    statistics_config=None,
    validation_type="NONE",
    expectations=[],
    event_time=None,
)

Create a feature group metadata object.

Lazy

This method is lazy and does not persist any metadata or feature data in the feature store on its own. To persist the feature group and save feature data along the metadata in the feature store, call the save() method with a DataFrame.

Arguments

  • name str: Name of the feature group to create.
  • version Optional[int]: Version of the feature group to retrieve, defaults to None and will create the feature group with incremented version from the last version in the feature store.
  • description Optional[str]: A string describing the contents of the feature group to improve discoverability for Data Scientists, defaults to empty string "".
  • online_enabled Optional[bool]: Define whether the feature group should be made available also in the online feature store for low latency access, defaults to False.
  • time_travel_format Optional[str]: Format used for time travel, defaults to "HUDI".
  • partition_key Optional[List[str]]: A list of feature names to be used as partition key when writing the feature data to the offline storage, defaults to empty list [].
  • primary_key Optional[List[str]]: A list of feature names to be used as primary key for the feature group. This primary key can be a composite key of multiple features and will be used as joining key, if not specified otherwise. Defaults to empty list [], and the feature group won't have any primary key.
  • hudi_precombine_key Optional[str]: A feature name to be used as a precombine key for the "HUDI" feature group. Defaults to None. If feature group has time travel format "HUDI" and hudi precombine key was not specified then the first primary key of the feature group will be used as hudi precombine key.
  • features Optional[List[hsfs.feature.Feature]]: Optionally, define the schema of the feature group manually as a list of Feature objects. Defaults to empty list [] and will use the schema information of the DataFrame provided in the save method.
  • statistics_config Optional[Union[hsfs.StatisticsConfig, bool, dict]]: A configuration object, or a dictionary with keys "enabled" to generally enable descriptive statistics computation for this feature group, "correlations" to turn on feature correlation computation, "histograms" to compute feature value frequencies and "exact_uniqueness" to compute uniqueness, distinctness and entropy. The values should be booleans indicating the setting. To fully turn off statistics computation pass statistics_config=False. Defaults to None and will compute only descriptive statistics.
  • validation_type Optional[str]: Optionally, set the validation type to one of "NONE", "STRICT", "WARNING", "ALL". Determines the mode in which data validation is applied on ingested or already existing feature group data.
  • expectations Optional[List[hsfs.expectation.Expectation]]: Optionally, a list of expectations to be attached to the feature group. The expectations list contains Expectation metadata objects which can be retrieved with the get_expectation() and get_expectations() functions.
  • event_time Optional[str]: Optionally, provide the name of the feature containing the event time for the features in this feature group. If event_time is set the feature group can be used for point-in-time joins. Defaults to None.

Returns

FeatureGroup. The feature group metadata object.

Retrieval#

[source]

get_feature_group#

FeatureStore.get_feature_group(name, version=None)

Get a feature group entity from the feature store.

Getting a feature group from the Feature Store means getting its metadata handle so you can subsequently read the data into a Spark or Pandas DataFrame or use the Query-API to perform joins between feature groups.

Arguments

  • name str: Name of the feature group to get.
  • version Optional[int]: Version of the feature group to retrieve, defaults to None and will return the version=1.

Returns

FeatureGroup: The feature group metadata object.

Raises

  • RestAPIError: If unable to retrieve feature group from the feature store.

Properties#

[source]

avro_schema#

Avro schema representation of the feature group.

[source]

created#

Timestamp when the feature group was created.

[source]

creator#

Username of the creator.

[source]

description#

Description of the feature group contents.

[source]

event_time#

Event time feature in the feature group.

[source]

expectations_names#

The names of expectations attached to this feature group.

[source]

feature_store_id#

[source]

feature_store_name#

Name of the feature store in which the feature group is located.

[source]

features#

Schema information.

[source]

hudi_precombine_key#

Feature name that is the hudi precombine key.

[source]

id#

Feature group id.

[source]

location#

[source]

name#

Name of the feature group.

[source]

online_enabled#

Setting if the feature group is available in online storage.

[source]

partition_key#

List of features building the partition key.

[source]

primary_key#

List of features building the primary key.

[source]

statistics#

Get the latest computed statistics for the feature group.

[source]

statistics_config#

Statistics configuration object defining the settings for statistics computation of the feature group.

[source]

time_travel_format#

Setting of the feature group time travel format.

[source]

validation_type#

Validation type, one of "STRICT", "WARNING", "ALL", "NONE".

[source]

version#

Version number of the feature group.

Methods#

[source]

add_tag#

FeatureGroup.add_tag(name, value)

Attach a tag to a feature group.

A tag consists of a pair. Tag names are unique identifiers across the whole cluster. The value of a tag can be any valid json - primitives, arrays or json objects.

Arguments

  • name str: Name of the tag to be added.
  • value: Value of the tag to be added.

Raises

RestAPIError in case the backend fails to add the tag.

[source]

append_features#

FeatureGroup.append_features(features)

Append features to the schema of the feature group.

Safe append

This method appends the features to the feature group description safely. In case of failure your local metadata object will contain the correct schema.

It is only possible to append features to a feature group. Removing features is considered a breaking change.

Arguments

  • features Union[hsfs.feature.Feature, List[hsfs.feature.Feature]]: Feature or list. A feature object or list thereof to append to the schema of the feature group.

Returns

FeatureGroup. The updated feature group object.

[source]

as_of#

FeatureGroup.as_of(wallclock_time)

Get Query object to retrieve all features of the group at a point in the past.

This method selects all features in the feature group and returns a Query object at the specified point in time. This can then either be read into a Dataframe or used further to perform joins or construct a training dataset.

Arguments

  • wallclock_time: Datetime string. The String should be formatted in one of the following formats %Y%m%d, %Y%m%d%H, %Y%m%d%H%M, or %Y%m%d%H%M%S.

Returns

Query. The query object with the applied time travel condition.

[source]

attach_expectation#

FeatureGroup.attach_expectation(expectation)

Attach a feature group expectation. If feature group validation is not already enabled, it will be enabled and set to the stricter setting.

Arguments

  • name: The expectation name.

Returns

Expectation. The expectation metadata object.

[source]

commit_delete_record#

FeatureGroup.commit_delete_record(delete_df, write_options={})

Drops records present in the provided DataFrame and commits it as update to this Feature group. This method can only be used on time travel enabled feature groups

Arguments

  • delete_df pyspark.sql.DataFrame: dataFrame containing records to be deleted.
  • write_options Optional[Dict[Any, Any]]: User provided write options. Defaults to {}.

Raises

RestAPIError.

[source]

commit_details#

FeatureGroup.commit_details(wallclock_time=None, limit=None)

Retrieves commit timeline for this feature group. This method can only be used on time travel enabled feature groups

Arguments

  • wallclock_time Optional[str]: Commit details as of specific point in time. Defaults to None.
  • limit Optional[int]: Number of commits to retrieve. Defaults to None.

Returns

Dict[str, Dict[str, str]]. Dictionary object of commit metadata timeline, where Key is commit id and value is Dict[str, str] with key value pairs of date committed on, number of rows updated, inserted and deleted.

Raises

RestAPIError. FeatureStoreException. If the feature group does not have HUDI time travel format

[source]

compute_statistics#

FeatureGroup.compute_statistics(wallclock_time=None)

Recompute the statistics for the feature group and save them to the feature store.

Statistics are only computed for data in the offline storage of the feature group.

Arguments

  • wallclock_time Optional[str]: Date string in the format of "YYYYMMDD" or "YYYYMMDDhhmmss". Only valid if feature group is time travel enabled. If specified will recompute statistics on feature group as of specific point in time. If not specified then will compute statistics as of most recent time of this feature group. Defaults to None.

Returns

Statistics. The statistics metadata object.

Raises

RestAPIError. Unable to persist the statistics.

[source]

delete#

FeatureGroup.delete()

Drop the entire feature group along with its feature data.

Potentially dangerous operation

This operation drops all metadata associated with this version of the feature group and all the feature data in offline and online storage associated with it.

Raises

RestAPIError.

[source]

delete_tag#

FeatureGroup.delete_tag(name)

Delete a tag attached to a feature group.

Arguments

  • name str: Name of the tag to be removed.

Raises

RestAPIError in case the backend fails to delete the tag.

[source]

detach_expectation#

FeatureGroup.detach_expectation(expectation)

Remove an expectation from a feature group.

Arguments

  • name: The expectation name.

Returns

Expectation. The expectation metadata object.

[source]

filter#

FeatureGroup.filter(f)

Apply filter to the feature group.

Selects all features and returns the resulting Query with the applied filter.

from hsfs.feature import Feature

fg.filter(Feature("weekly_sales") > 1000)

If you are planning to join the filtered feature group later on with another feature group, make sure to select the filtered feature explicitly from the respective feature group: 

fg.filter(fg.feature1 == 1).show(10)

Composite filters require parenthesis: 

fg.filter((fg.feature1 == 1) | (fg.feature2 >= 2))

Arguments

  • f Union[hsfs.constructor.filter.Filter, hsfs.constructor.filter.Logic]: Filter object.

Returns

Query. The query object with the applied filter.

[source]

from_response_json#

FeatureGroup.from_response_json(json_dict)

[source]

get_complex_features#

FeatureGroup.get_complex_features()

Returns the names of all features with a complex data type in this feature group.

[source]

get_expectation#

FeatureGroup.get_expectation(name)

Get attached expectation by name for this feature group. Name is unique across a feature store.

Arguments

  • name str: The expectation name.

Returns

Expectation. The expectation metadata object.

[source]

get_expectations#

FeatureGroup.get_expectations()

Get all feature group expectations.

Arguments

  • name: The expectation name.

Returns

Expectation. A list of expectation metadata objects.

[source]

get_feature#

FeatureGroup.get_feature(name)

Retrieve a Feature object from the schema of the feature group.

There are several ways to access features of a feature group:

fg.feature1
fg["feature1"]
fg.get_feature("feature1")

Note

Attribute access to features works only for non-reserved names. For example features named id or name will not be accessible via fg.name, instead this will return the name of the feature group itself. Fall back on using the get_feature method.

Args: name (str): [description]

Returns: [type]: [description]

[source]

get_statistics#

FeatureGroup.get_statistics(commit_time=None)

Returns the statistics for this feature group at a specific time.

If commit_time is None, the most recent statistics are returned.

Arguments

  • commit_time Optional[str]: Commit time in the format YYYYMMDDhhmmss, defaults to None.

Returns

Statistics. Statistics object.

Raises

RestAPIError.

[source]

get_tag#

FeatureGroup.get_tag(name)

Get the tags of a feature group.

Arguments

  • name str: Name of the tag to get.

Returns

tag value

Raises

RestAPIError in case the backend fails to retrieve the tag.

[source]

get_tags#

FeatureGroup.get_tags()

Retrieves all tags attached to a feature group.

Returns

Dict[str, obj] of tags.

Raises

RestAPIError in case the backend fails to retrieve the tags.

[source]

get_validations#

FeatureGroup.get_validations(validation_time=None, commit_time=None)

Get feature group data validation results based on the attached expectations.

Arguments

  • validation_time: The data validation time, when the data validation started. commit_time: The commit time of a time travel enabled feature group.

Returns

FeatureGroupValidation. The feature group validation metadata object.

[source]

insert#

FeatureGroup.insert(
    features, overwrite=False, operation="upsert", storage=None, write_options={}
)

Insert data from a dataframe into the feature group.

Incrementally insert data to a feature group or overwrite all data contained in the feature group. By default, the data is inserted into the offline storage as well as the online storage if the feature group is online_enabled=True. To insert only into the online storage, set storage="online", or oppositely storage="offline".

The features dataframe can be a Spark DataFrame or RDD, a Pandas DataFrame, or a two-dimensional Numpy array or a two-dimensional Python nested list.

If statistics are enabled, statistics are recomputed for the entire feature group.

If feature group's time travel format is HUDI then operation argument can be either insert or upsert.

Upsert new feature data with time travel format HUDI:

fs = conn.get_feature_store();
fg = fs.get_feature_group("example_feature_group", 1)
upsert_df = ...
fg.insert(upsert_df)

Arguments

  • features Union[pandas.DataFrame, pyspark.sql.DataFrame, pyspark.RDD, numpy.ndarray, List[list]]: DataFrame, RDD, Ndarray, list. Features to be saved.
  • overwrite Optional[bool]: Drop all data in the feature group before inserting new data. This does not affect metadata, defaults to False.
  • operation Optional[str]: Apache Hudi operation type "insert" or "upsert". Defaults to "upsert".
  • storage Optional[str]: Overwrite default behaviour, write to offline storage only with "offline" or online only with "online", defaults to None.
  • write_options Optional[Dict[Any, Any]]: Additional write options as key-value pairs, defaults to {}. When using the hive engine, write_options can contain the following entries:
    • key spark and value an object of type hsfs.core.job_configuration.JobConfiguration to configure the Hopsworks Job used to write data into the feature group.
    • key wait_for_job and value True or False to configure whether or not to the insert call should return only after the Hopsworks Job has finished. By default it waits.

Returns

FeatureGroup. Updated feature group metadata object.

[source]

insert_stream#

FeatureGroup.insert_stream(
    features,
    query_name=None,
    output_mode="append",
    await_termination=False,
    timeout=None,
    checkpoint_dir=None,
    write_options={},
)

Ingest a Spark Structured Streaming Dataframe to the online feature store.

This method creates a long running Spark Streaming Query, you can control the termination of the query through the arguments.

It is possible to stop the returned query with the .stop() and check its status with .isActive.

To get a list of all active queries, use:

sqm = spark.streams
# get the list of active streaming queries
[q.name for q in sqm.active]

Arguments

  • features pyspark.sql.DataFrame: Features in Streaming Dataframe to be saved.
  • query_name Optional[str]: It is possible to optionally specify a name for the query to make it easier to recognise in the Spark UI. Defaults to None.
  • output_mode Optional[str]: Specifies how data of a streaming DataFrame/Dataset is written to a streaming sink. (1) "append": Only the new rows in the streaming DataFrame/Dataset will be written to the sink. (2) "complete": All the rows in the streaming DataFrame/Dataset will be written to the sink every time there is some update. (3) "update": only the rows that were updated in the streaming DataFrame/Dataset will be written to the sink every time there are some updates. If the query doesn’t contain aggregations, it will be equivalent to append mode. Defaults to "append".
  • await_termination Optional[bool]: Waits for the termination of this query, either by query.stop() or by an exception. If the query has terminated with an exception, then the exception will be thrown. If timeout is set, it returns whether the query has terminated or not within the timeout seconds. Defaults to False.
  • timeout Optional[int]: Only relevant in combination with await_termination=True. Defaults to None.
  • checkpoint_dir Optional[str]: Checkpoint directory location. This will be used to as a reference to from where to resume the streaming job. If None then hsfs will construct as "insert_stream_" + online_topic_name. Defaults to None.
  • write_options Optional[Dict[Any, Any]]: Additional write options for Spark as key-value pairs. Defaults to {}.

Returns

StreamingQuery: Spark Structured Streaming Query object.

[source]

json#

FeatureGroup.json()

[source]

read#

FeatureGroup.read(wallclock_time=None, online=False, dataframe_type="default", read_options={})

Read the feature group into a dataframe.

Reads the feature group by default from the offline storage as Spark DataFrame on Hopsworks and Databricks, and as Pandas dataframe on AWS Sagemaker and pure Python environments.

Set online to True to read from the online storage, or change dataframe_type to read as a different format.

Read feature group as of latest state:

fs = connection.get_feature_store();
fg = fs.get_feature_group("example_feature_group", 1)
fg.read()

Read feature group as of specific point in time:

fs = connection.get_feature_store();
fg = fs.get_feature_group("example_feature_group", 1)
fg.read("2020-10-20 07:34:11")

Arguments

  • wallclock_time Optional[str]: Date string in the format of "YYYYMMDD" or "YYYYMMDDhhmmss". If Specified will retrieve feature group as of specific point in time. If not specified will return as of most recent time. Defaults to None.
  • online Optional[bool]: bool, optional. If True read from online feature store, defaults to False.
  • dataframe_type Optional[str]: str, optional. Possible values are "default", "spark", "pandas", "numpy" or "python", defaults to "default".
  • read_options Optional[dict]: Additional read options as key/value pairs, defaults to {}.

Returns

DataFrame: The spark dataframe containing the feature data. pyspark.DataFrame. A Spark DataFrame. pandas.DataFrame. A Pandas DataFrame. numpy.ndarray. A two-dimensional Numpy array. list. A two-dimensional Python list.

Raises

RestAPIError. No data is available for feature group with this commit date, If time travel enabled.

[source]

read_changes#

FeatureGroup.read_changes(start_wallclock_time, end_wallclock_time, read_options={})

Reads updates of this feature that occurred between specified points in time.

This function only works on feature groups with HUDI time travel format.

Reading commits incrementally between specified points in time:

fs = connection.get_feature_store();
fg = fs.get_feature_group("example_feature_group", 1)
fg.read_changes("2020-10-20 07:31:38", "2020-10-20 07:34:11").show()

Arguments

  • start_wallclock_time str: Date string in the format of "YYYYMMDD" or "YYYYMMDDhhmmss".
  • end_wallclock_time str: Date string in the format of "YYYYMMDD" or "YYYYMMDDhhmmss".
  • read_options Optional[dict]: User provided read options. Defaults to {}.

Returns

DataFrame. The spark dataframe containing the incremental changes of feature data.

Raises

RestAPIError. No data is available for feature group with this commit date. FeatureStoreException. If the feature group does not have HUDI time travel format

[source]

save#

FeatureGroup.save(features, write_options={})

Persist the metadata and materialize the feature group to the feature store.

Calling save creates the metadata for the feature group in the feature store and writes the specified features dataframe as feature group to the online/offline feature store as specified.

By default, this writes the feature group to the offline storage, and if online_enabled for the feature group, also to the online feature store.

The features dataframe can be a Spark DataFrame or RDD, a Pandas DataFrame, or a two-dimensional Numpy array or a two-dimensional Python nested list.

Arguments

  • features Union[pandas.DataFrame, pyspark.sql.DataFrame, pyspark.RDD, numpy.ndarray, List[list]]: Query, DataFrame, RDD, Ndarray, list. Features to be saved.
  • write_options Optional[Dict[Any, Any]]: Additional write options as key-value pairs, defaults to {}. When using the hive engine, write_options can contain the following entries:
    • key spark and value an object of type hsfs.core.job_configuration.JobConfiguration to configure the Hopsworks Job used to write data into the feature group.
    • key wait_for_job and value True or False to configure whether or not to the save call should return only after the Hopsworks Job has finished. By default it waits.

Returns

Job: When using the hive engine, it returns the Hopsworks Job that was launched to ingest the feature group data.

Raises

RestAPIError. Unable to create feature group.

[source]

select#

FeatureGroup.select(features=[])

Select a subset of features of the feature group and return a query object.

The query can be used to construct joins of feature groups or create a training dataset with a subset of features of the feature group.

Arguments

  • features List[Union[str, hsfs.feature.Feature]]: list, optional. A list of Feature objects or feature names as strings to be selected, defaults to [].

Returns

Query: A query object with the selected features of the feature group.

[source]

select_all#

FeatureGroup.select_all()

Select all features in the feature group and return a query object.

The query can be used to construct joins of feature groups or create a training dataset immediately.

Returns

Query. A query object with all features of the feature group.

[source]

select_except#

FeatureGroup.select_except(features=[])

Select all features of the feature group except a few and return a query object.

The query can be used to construct joins of feature groups or create a training dataset with a subset of features of the feature group.

Arguments

  • features List[Union[str, hsfs.feature.Feature]]: list, optional. A list of Feature objects or feature names as strings to be excluded from the selection. Defaults to [], selecting all features.

Returns

Query: A query object with the selected features of the feature group.

[source]

show#

FeatureGroup.show(n, online=False)

Show the first n rows of the feature group.

Arguments

  • n int: int. Number of rows to show.
  • online Optional[bool]: bool, optional. If True read from online feature store, defaults to False.

[source]

to_dict#

FeatureGroup.to_dict()

[source]

update_description#

FeatureGroup.update_description(description)

Update the description of the feature group.

Safe update

This method updates the feature group description safely. In case of failure your local metadata object will keep the old description.

Arguments

  • description str: New description string.

Returns

FeatureGroup. The updated feature group object.

[source]

update_feature_description#

FeatureGroup.update_feature_description(feature_name, description)

Update the description of a single feature in this feature group.

Safe update

This method updates the feature description safely. In case of failure your local metadata object will keep the old description.

Arguments

  • feature_name str: Name of the feature to be updated.
  • description str: New description string.

Returns

FeatureGroup. The updated feature group object.

[source]

update_features#

FeatureGroup.update_features(features)

Update a single feature in this feature group.

Currently it's only supported to update the description of a feature.

Unsafe update

Note that if you use an existing Feature object of the schema in the feature group metadata object, this might leave your metadata object in a corrupted state if the update fails.

Arguments

  • features Union[hsfs.feature.Feature, List[hsfs.feature.Feature]]: Feature or list of features. A feature object or list thereof to be updated.

Returns

FeatureGroup. The updated feature group object.

[source]

update_from_response_json#

FeatureGroup.update_from_response_json(json_dict)

[source]

update_statistics_config#

FeatureGroup.update_statistics_config()

Update the statistics configuration of the feature group.

Change the statistics_config object and persist the changes by calling this method.

Returns

FeatureGroup. The updated metadata object of the feature group.

Raises

RestAPIError.

[source]

validate#

FeatureGroup.validate(dataframe=None, log_activity=False)

Run validation based on the attached expectations

Arguments

  • dataframe Optional[pyspark.sql.DataFrame]: The PySpark dataframe to run the data validation expectations against.
  • log_activity: Whether to log the validation as a feature group activity. If a dataframe is not provided (None), the validation will be logged as a feature store activity.

Returns

FeatureGroupValidation. The feature group validation metadata object.