Legacy proto-based Types for Google Cloud Bigquery v2 API

WARNING: These types are provided for backward compatibility only, and are not maintainedanymore. They might also differ from the types uspported on the backend. It istherefore strongly advised to migrate to the types found inTypes for Google Cloud Bigquery v2 API.

Also see the3.0.0 Migration Guide for more information.

class google.cloud.bigquery_v2.types.DeleteModelRequest(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

project_id()

Required. Project ID of the model to delete.

dataset_id()

Required. Dataset ID of the model to delete.

model_id()

Required. Model ID of the model to delete.

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

class google.cloud.bigquery_v2.types.EncryptionConfiguration(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

kms_key_name()

Optional. Describes the Cloud KMS encryptionkey that will be used to protect destinationBigQuery table. The BigQuery Service Accountassociated with your project requires access tothis encryption key.

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

class google.cloud.bigquery_v2.types.GetModelRequest(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

project_id()

Required. Project ID of the requested model.

dataset_id()

Required. Dataset ID of the requested model.

model_id()

Required. Model ID of the requested model.

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

class google.cloud.bigquery_v2.types.ListModelsRequest(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

project_id()

Required. Project ID of the models to list.

dataset_id()

Required. Dataset ID of the models to list.

max_results()

The maximum number of results to return in asingle response page. Leverage the page tokensto iterate through the entire collection.

page_token()

Page token, returned by a previous call torequest the next page of results

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

class google.cloud.bigquery_v2.types.ListModelsResponse(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

models()

Models in the requested dataset. Only the following fieldsare populated: model_reference, model_type, creation_time,last_modified_time and labels.

  • Type

    Sequence[google.cloud.bigquery_v2.types.Model]

next_page_token()

A token to request the next page of results.

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

class google.cloud.bigquery_v2.types.Model(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

etag()

Output only. A hash of this resource.

model_reference()

Required. Unique identifier for this model.

  • Type

    google.cloud.bigquery_v2.types.ModelReference

creation_time()

Output only. The time when this model wascreated, in millisecs since the epoch.

last_modified_time()

Output only. The time when this model waslast modified, in millisecs since the epoch.

description()

Optional. A user-friendly description of thismodel.

friendly_name()

Optional. A descriptive name for this model.

labels()

The labels associated with this model. Youcan use these to organize and group your models.Label keys and values can be no longer than 63characters, can only contain lowercase letters,numeric characters, underscores and dashes.International characters are allowed. Labelvalues are optional. Label keys must start witha letter and each label in the list must have adifferent key.

expiration_time()

Optional. The time when this model expires,in milliseconds since the epoch. If not present,the model will persist indefinitely. Expiredmodels will be deleted and their storagereclaimed. The defaultTableExpirationMsproperty of the encapsulating dataset can beused to set a default expirationTime on newlycreated models.

location()

Output only. The geographic location wherethe model resides. This value is inherited fromthe dataset.

encryption_configuration()

Custom encryption configuration (e.g., CloudKMS keys). This shows the encryptionconfiguration of the model data while stored inBigQuery storage. This field can be used withPatchModel to update encryption key for analready encrypted model.

  • Type

    google.cloud.bigquery_v2.types.EncryptionConfiguration

model_type()

Output only. Type of the model resource.

  • Type

    google.cloud.bigquery_v2.types.Model.ModelType

training_runs()

Output only. Information for all training runs in increasingorder of start_time.

  • Type

    Sequence[google.cloud.bigquery_v2.types.Model.TrainingRun]

feature_columns()

Output only. Input feature columns that wereused to train this model.

  • Type

    Sequence[google.cloud.bigquery_v2.types.StandardSqlField]

label_columns()

Output only. Label columns that were used to train thismodel. The output of the model will have a predicted_prefix to these columns.

  • Type

    Sequence[google.cloud.bigquery_v2.types.StandardSqlField]

best_trial_id()

The best trial_id across all training runs.

class AggregateClassificationMetrics(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

Aggregate metrics for classification/classifier models. Formulti-class models, the metrics are either macro-averaged ormicro-averaged. When macro-averaged, the metrics are calculatedfor each label and then an unweighted average is taken of thosevalues. When micro-averaged, the metric is calculated globallyby counting the total number of correctly predicted rows.

precision()

Precision is the fraction of actual positivepredictions that had positive actual labels. Formulticlass this is a macro-averaged metrictreating each class as a binary classifier.

recall()

Recall is the fraction of actual positivelabels that were given a positive prediction.For multiclass this is a macro-averaged metric.

accuracy()

Accuracy is the fraction of predictions giventhe correct label. For multiclass this is amicro-averaged metric.

threshold()

Threshold at which the metrics are computed.For binary classification models this is thepositive class threshold. For multi-classclassfication models this is the confidencethreshold.

f1_score()

The F1 score is an average of recall andprecision. For multiclass this is amacro-averaged metric.

log_loss()

Logarithmic Loss. For multiclass this is amacro-averaged metric.

roc_auc()

Area Under a ROC Curve. For multiclass thisis a macro-averaged metric.

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

class ArimaFittingMetrics(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

ARIMA model fitting metrics.

log_likelihood()

Log-likelihood.

aic()

AIC.

variance()

Variance.

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

class ArimaForecastingMetrics(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

Model evaluation metrics for ARIMA forecasting models.

non_seasonal_order()

Non-seasonal order.

  • Type

    Sequence[google.cloud.bigquery_v2.types.Model.ArimaOrder]

arima_fitting_metrics()

Arima model fitting metrics.

  • Type

    Sequence[google.cloud.bigquery_v2.types.Model.ArimaFittingMetrics]

seasonal_periods()

Seasonal periods. Repeated because multipleperiods are supported for one time series.

  • Type

    Sequence[google.cloud.bigquery_v2.types.Model.SeasonalPeriod.SeasonalPeriodType]

has_drift()

Whether Arima model fitted with drift or not.It is always false when d is not 1.

time_series_id()

Id to differentiate different time series forthe large-scale case.

  • Type

    Sequence[str]

arima_single_model_forecasting_metrics()

Repeated as there can be many metric sets(one for each model) in auto-arima and thelarge-scale case.

  • Type

    Sequence[google.cloud.bigquery_v2.types.Model.ArimaForecastingMetrics.ArimaSingleModelForecastingMetrics]

class ArimaSingleModelForecastingMetrics(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

Model evaluation metrics for a single ARIMA forecastingmodel.

non_seasonal_order()

Non-seasonal order.

  • Type

    google.cloud.bigquery_v2.types.Model.ArimaOrder

arima_fitting_metrics()

Arima fitting metrics.

  • Type

    google.cloud.bigquery_v2.types.Model.ArimaFittingMetrics

has_drift()

Is arima model fitted with drift or not. Itis always false when d is not 1.

time_series_id()

The time_series_id value for this time series. It will beone of the unique values from the time_series_id_columnspecified during ARIMA model training. Only present whentime_series_id_column training option was used.

time_series_ids()

The tuple of time_series_ids identifying this time series.It will be one of the unique tuples of values present in thetime_series_id_columns specified during ARIMA modeltraining. Only present when time_series_id_columns trainingoption was used and the order of values here are same as theorder of time_series_id_columns.

  • Type

    Sequence[str]

seasonal_periods()

Seasonal periods. Repeated because multipleperiods are supported for one time series.

  • Type

    Sequence[google.cloud.bigquery_v2.types.Model.SeasonalPeriod.SeasonalPeriodType]

has_holiday_effect()

If true, holiday_effect is a part of time seriesdecomposition result.

has_spikes_and_dips()

If true, spikes_and_dips is a part of time seriesdecomposition result.

has_step_changes()

If true, step_changes is a part of time series decompositionresult.

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

class ArimaOrder(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

Arima order, can be used for both non-seasonal and seasonalparts.

p()

Order of the autoregressive part.

d()

Order of the differencing part.

q()

Order of the moving-average part.

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

class BinaryClassificationMetrics(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

Evaluation metrics for binary classification/classifiermodels.

aggregate_classification_metrics()

Aggregate classification metrics.

  • Type

    google.cloud.bigquery_v2.types.Model.AggregateClassificationMetrics

binary_confusion_matrix_list()

Binary confusion matrix at multiplethresholds.

  • Type

    Sequence[google.cloud.bigquery_v2.types.Model.BinaryClassificationMetrics.BinaryConfusionMatrix]

positive_label()

Label representing the positive class.

negative_label()

Label representing the negative class.

class BinaryConfusionMatrix(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

Confusion matrix for binary classification models.

positive_class_threshold()

Threshold value used when computing each ofthe following metric.

true_positives()

Number of true samples predicted as true.

false_positives()

Number of false samples predicted as true.

true_negatives()

Number of true samples predicted as false.

false_negatives()

Number of false samples predicted as false.

precision()

The fraction of actual positive predictionsthat had positive actual labels.

recall()

The fraction of actual positive labels thatwere given a positive prediction.

f1_score()

The equally weighted average of recall andprecision.

accuracy()

The fraction of predictions given the correctlabel.

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

class ClusteringMetrics(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

Evaluation metrics for clustering models.

davies_bouldin_index()

Davies-Bouldin index.

mean_squared_distance()

Mean of squared distances between each sampleto its cluster centroid.

clusters()

Information for all clusters.

  • Type

    Sequence[google.cloud.bigquery_v2.types.Model.ClusteringMetrics.Cluster]

class Cluster(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

Message containing the information about one cluster.

centroid_id()

Centroid id.

feature_values()

Values of highly variant features for thiscluster.

  • Type

    Sequence[google.cloud.bigquery_v2.types.Model.ClusteringMetrics.Cluster.FeatureValue]

count()

Count of training data rows that wereassigned to this cluster.

class FeatureValue(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

Representative value of a single feature within the cluster.

This message hasoneof fields (mutually exclusive fields).For each oneof, at most one member field can be set at the same time.Setting any member of the oneof automatically clears all othermembers.

feature_column()

The feature column name.

numerical_value()

The numerical feature value. This is thecentroid value for this feature.

This field is a member ofoneofvalue.

categorical_value()

The categorical feature value.

This field is a member ofoneofvalue.

  • Type

    google.cloud.bigquery_v2.types.Model.ClusteringMetrics.Cluster.FeatureValue.CategoricalValue

class CategoricalValue(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

Representative value of a categorical feature.

category_counts()

Counts of all categories for the categorical feature. Ifthere are more than ten categories, we return top ten (bycount) and return one more CategoryCount with category“OTHER” and count as aggregate counts of remainingcategories.

  • Type

    Sequence[google.cloud.bigquery_v2.types.Model.ClusteringMetrics.Cluster.FeatureValue.CategoricalValue.CategoryCount]

class CategoryCount(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

Represents the count of a single category within the cluster.

category()

The name of category.

count()

The count of training samples matching thecategory within the cluster.

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

class DataFrequency(value)

Bases:proto.enums.Enum

Type of supported data frequency for time series forecastingmodels.

AUTO_FREQUENCY( = )

DAILY( = )

DATA_FREQUENCY_UNSPECIFIED( = )

HOURLY( = )

MONTHLY( = )

PER_MINUTE( = )

QUARTERLY( = )

WEEKLY( = )

YEARLY( = )

class DataSplitMethod(value)

Bases:proto.enums.Enum

Indicates the method to split input data into multipletables.

AUTO_SPLIT( = )

CUSTOM( = )

DATA_SPLIT_METHOD_UNSPECIFIED( = )

NO_SPLIT( = )

RANDOM( = )

SEQUENTIAL( = )

class DataSplitResult(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

Data split result. This contains references to the trainingand evaluation data tables that were used to train the model.

training_table()

Table reference of the training data aftersplit.

  • Type

    google.cloud.bigquery_v2.types.TableReference

evaluation_table()

Table reference of the evaluation data aftersplit.

  • Type

    google.cloud.bigquery_v2.types.TableReference

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

class DistanceType(value)

Bases:proto.enums.Enum

Distance metric used to compute the distance between twopoints.

COSINE( = )

DISTANCE_TYPE_UNSPECIFIED( = )

EUCLIDEAN( = )

class EvaluationMetrics(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

Evaluation metrics of a model. These are either computed onall training data or just the eval data based on whether evaldata was used during training. These are not present forimported models.

This message hasoneof fields (mutually exclusive fields).For each oneof, at most one member field can be set at the same time.Setting any member of the oneof automatically clears all othermembers.

regression_metrics()

Populated for regression models and explicitfeedback type matrix factorization models.

This field is a member ofoneofmetrics.

  • Type

    google.cloud.bigquery_v2.types.Model.RegressionMetrics

binary_classification_metrics()

Populated for binaryclassification/classifier models.

This field is a member ofoneofmetrics.

  • Type

    google.cloud.bigquery_v2.types.Model.BinaryClassificationMetrics

multi_class_classification_metrics()

Populated for multi-classclassification/classifier models.

This field is a member ofoneofmetrics.

  • Type

    google.cloud.bigquery_v2.types.Model.MultiClassClassificationMetrics

clustering_metrics()

Populated for clustering models.

This field is a member ofoneofmetrics.

  • Type

    google.cloud.bigquery_v2.types.Model.ClusteringMetrics

ranking_metrics()

Populated for implicit feedback type matrixfactorization models.

This field is a member ofoneofmetrics.

  • Type

    google.cloud.bigquery_v2.types.Model.RankingMetrics

arima_forecasting_metrics()

Populated for ARIMA models.

This field is a member ofoneofmetrics.

  • Type

    google.cloud.bigquery_v2.types.Model.ArimaForecastingMetrics

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

class FeedbackType(value)

Bases:proto.enums.Enum

Indicates the training algorithm to use for matrixfactorization models.

EXPLICIT( = )

FEEDBACK_TYPE_UNSPECIFIED( = )

IMPLICIT( = )

class GlobalExplanation(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

Global explanations containing the top most importantfeatures after training.

explanations()

A list of the top global explanations. Sortedby absolute value of attribution in descendingorder.

  • Type

    Sequence[google.cloud.bigquery_v2.types.Model.GlobalExplanation.Explanation]

class_label()

Class label for this set of globalexplanations. Will be empty/null for binarylogistic and linear regression models. Sortedalphabetically in descending order.

class Explanation(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

Explanation for a single feature.

feature_name()

Full name of the feature. For non-numerical features, willbe formatted like <column_name>.<encoded_feature_name>.Overall size of feature name will always be truncated tofirst 120 characters.

attribution()

Attribution of feature.

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

class HolidayRegion(value)

Bases:proto.enums.Enum

Type of supported holiday regions for time series forecastingmodels.

AE( = )

AR( = )

AT( = )

AU( = )

BE( = 1 )

BR( = 1 )

CA( = 1 )

CH( = 1 )

CL( = 1 )

CN( = 1 )

CO( = 1 )

CS( = 1 )

CZ( = 1 )

DE( = 1 )

DK( = 2 )

DZ( = 2 )

EC( = 2 )

EE( = 2 )

EG( = 2 )

EMEA( = )

ES( = 2 )

FI( = 2 )

FR( = 2 )

GB( = 2 )

GLOBAL( = )

GR( = 2 )

HK( = 3 )

HOLIDAY_REGION_UNSPECIFIED( = )

HU( = 3 )

ID( = 3 )

IE( = 3 )

IL( = 3 )

IN( = 3 )

IR( = 3 )

IT( = 3 )

JAPAC( = )

JP( = 3 )

KR( = 3 )

LAC( = )

LV( = 4 )

MA( = 4 )

MX( = 4 )

MY( = 4 )

NA( = )

NG( = 4 )

NL( = 4 )

NO( = 4 )

NZ( = 4 )

PE( = 4 )

PH( = 4 )

PK( = 5 )

PL( = 5 )

PT( = 5 )

RO( = 5 )

RS( = 5 )

RU( = 5 )

SA( = 5 )

SE( = 5 )

SG( = 5 )

SI( = 5 )

SK( = 6 )

TH( = 6 )

TR( = 6 )

TW( = 6 )

UA( = 6 )

US( = 6 )

VE( = 6 )

VN( = 6 )

ZA( = 6 )

class KmeansEnums(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

class KmeansInitializationMethod(value)

Bases:proto.enums.Enum

Indicates the method used to initialize the centroids forKMeans clustering algorithm.

CUSTOM( = )

KMEANS_INITIALIZATION_METHOD_UNSPECIFIED( = )

KMEANS_PLUS_PLUS( = )

RANDOM( = )

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

class LabelsEntry(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

class LearnRateStrategy(value)

Bases:proto.enums.Enum

Indicates the learning rate optimization strategy to use.

CONSTANT( = )

LEARN_RATE_STRATEGY_UNSPECIFIED( = )

LINE_SEARCH( = )

class LossType(value)

Bases:proto.enums.Enum

Loss metric to evaluate model training performance.

LOSS_TYPE_UNSPECIFIED( = )

MEAN_LOG_LOSS( = )

MEAN_SQUARED_LOSS( = )

class ModelType(value)

Bases:proto.enums.Enum

Indicates the type of the Model.

ARIMA( = 1 )

ARIMA_PLUS( = 1 )

AUTOML_CLASSIFIER( = 1 )

AUTOML_REGRESSOR( = 1 )

BOOSTED_TREE_CLASSIFIER( = 1 )

BOOSTED_TREE_REGRESSOR( = )

DNN_CLASSIFIER( = )

DNN_REGRESSOR( = )

KMEANS( = )

LINEAR_REGRESSION( = )

LOGISTIC_REGRESSION( = )

MATRIX_FACTORIZATION( = )

MODEL_TYPE_UNSPECIFIED( = )

TENSORFLOW( = )

class MultiClassClassificationMetrics(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

Evaluation metrics for multi-class classification/classifiermodels.

aggregate_classification_metrics()

Aggregate classification metrics.

  • Type

    google.cloud.bigquery_v2.types.Model.AggregateClassificationMetrics

confusion_matrix_list()

Confusion matrix at different thresholds.

  • Type

    Sequence[google.cloud.bigquery_v2.types.Model.MultiClassClassificationMetrics.ConfusionMatrix]

class ConfusionMatrix(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

Confusion matrix for multi-class classification models.

confidence_threshold()

Confidence threshold used when computing theentries of the confusion matrix.

rows()

One row per actual label.

  • Type

    Sequence[google.cloud.bigquery_v2.types.Model.MultiClassClassificationMetrics.ConfusionMatrix.Row]

class Entry(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

A single entry in the confusion matrix.

predicted_label()

The predicted label. For confidence_threshold > 0, we willalso add an entry indicating the number of items under theconfidence threshold.

item_count()

Number of items being predicted as thislabel.

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

class Row(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

A single row in the confusion matrix.

actual_label()

The original label of this row.

entries()

Info describing predicted label distribution.

  • Type

    Sequence[google.cloud.bigquery_v2.types.Model.MultiClassClassificationMetrics.ConfusionMatrix.Entry]

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

class OptimizationStrategy(value)

Bases:proto.enums.Enum

Indicates the optimization strategy used for training.

BATCH_GRADIENT_DESCENT( = )

NORMAL_EQUATION( = )

OPTIMIZATION_STRATEGY_UNSPECIFIED( = )

class RankingMetrics(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

Evaluation metrics used by weighted-ALS models specified byfeedback_type=implicit.

mean_average_precision()

Calculates a precision per user for all theitems by ranking them and then averages all theprecisions across all the users.

mean_squared_error()

Similar to the mean squared error computed inregression and explicit recommendation modelsexcept instead of computing the rating directly,the output from evaluate is computed against apreference which is 1 or 0 depending on if therating exists or not.

normalized_discounted_cumulative_gain()

A metric to determine the goodness of aranking calculated from the predicted confidenceby comparing it to an ideal rank measured by theoriginal ratings.

average_rank()

Determines the goodness of a ranking bycomputing the percentile rank from the predictedconfidence and dividing it by the original rank.

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

class RegressionMetrics(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

Evaluation metrics for regression and explicit feedback typematrix factorization models.

mean_absolute_error()

Mean absolute error.

mean_squared_error()

Mean squared error.

mean_squared_log_error()

Mean squared log error.

median_absolute_error()

Median absolute error.

r_squared()

R^2 score. This corresponds to r2_score in ML.EVALUATE.

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

class SeasonalPeriod(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

class SeasonalPeriodType(value)

Bases:proto.enums.Enum

DAILY( = )

MONTHLY( = )

NO_SEASONALITY( = )

QUARTERLY( = )

SEASONAL_PERIOD_TYPE_UNSPECIFIED( = )

WEEKLY( = )

YEARLY( = )

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

class TrainingRun(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

Information about a single training query run for the model.

training_options()

Options that were used for this training run,includes user specified and default options thatwere used.

  • Type

    google.cloud.bigquery_v2.types.Model.TrainingRun.TrainingOptions

start_time()

The start time of this training run.

results()

Output of each iteration run, results.size() <=max_iterations.

  • Type

    Sequence[google.cloud.bigquery_v2.types.Model.TrainingRun.IterationResult]

evaluation_metrics()

The evaluation metrics over training/evaldata that were computed at the end of training.

  • Type

    google.cloud.bigquery_v2.types.Model.EvaluationMetrics

data_split_result()

Data split result of the training run. Onlyset when the input data is actually split.

  • Type

    google.cloud.bigquery_v2.types.Model.DataSplitResult

global_explanations()

Global explanations for important features ofthe model. For multi-class models, there is oneentry for each label class. For other models,there is only one entry in the list.

  • Type

    Sequence[google.cloud.bigquery_v2.types.Model.GlobalExplanation]

class IterationResult(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

Information about a single iteration of the training run.

index()

Index of the iteration, 0 based.

duration_ms()

Time taken to run the iteration inmilliseconds.

training_loss()

Loss computed on the training data at the endof iteration.

eval_loss()

Loss computed on the eval data at the end ofiteration.

learn_rate()

Learn rate used for this iteration.

cluster_infos()

Information about top clusters for clusteringmodels.

  • Type

    Sequence[google.cloud.bigquery_v2.types.Model.TrainingRun.IterationResult.ClusterInfo]

arima_result()

  • Type

    google.cloud.bigquery_v2.types.Model.TrainingRun.IterationResult.ArimaResult

class ArimaResult(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

(Auto-)arima fitting result. Wrap everything in ArimaResultfor easier refactoring if we want to use model-specificiteration results.

arima_model_info()

This message is repeated because there aremultiple arima models fitted in auto-arima. Fornon-auto-arima model, its size is one.

  • Type

    Sequence[google.cloud.bigquery_v2.types.Model.TrainingRun.IterationResult.ArimaResult.ArimaModelInfo]

seasonal_periods()

Seasonal periods. Repeated because multipleperiods are supported for one time series.

  • Type

    Sequence[google.cloud.bigquery_v2.types.Model.SeasonalPeriod.SeasonalPeriodType]

class ArimaCoefficients(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

Arima coefficients.

auto_regressive_coefficients()

Auto-regressive coefficients, an array ofdouble.

moving_average_coefficients()

Moving-average coefficients, an array ofdouble.

intercept_coefficient()

Intercept coefficient, just a double not anarray.

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

class ArimaModelInfo(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

Arima model information.

non_seasonal_order()

Non-seasonal order.

  • Type

    google.cloud.bigquery_v2.types.Model.ArimaOrder

arima_coefficients()

Arima coefficients.

  • Type

    google.cloud.bigquery_v2.types.Model.TrainingRun.IterationResult.ArimaResult.ArimaCoefficients

arima_fitting_metrics()

Arima fitting metrics.

  • Type

    google.cloud.bigquery_v2.types.Model.ArimaFittingMetrics

has_drift()

Whether Arima model fitted with drift or not.It is always false when d is not 1.

time_series_id()

The time_series_id value for this time series. It will beone of the unique values from the time_series_id_columnspecified during ARIMA model training. Only present whentime_series_id_column training option was used.

time_series_ids()

The tuple of time_series_ids identifying this time series.It will be one of the unique tuples of values present in thetime_series_id_columns specified during ARIMA modeltraining. Only present when time_series_id_columns trainingoption was used and the order of values here are same as theorder of time_series_id_columns.

  • Type

    Sequence[str]

seasonal_periods()

Seasonal periods. Repeated because multipleperiods are supported for one time series.

  • Type

    Sequence[google.cloud.bigquery_v2.types.Model.SeasonalPeriod.SeasonalPeriodType]

has_holiday_effect()

If true, holiday_effect is a part of time seriesdecomposition result.

has_spikes_and_dips()

If true, spikes_and_dips is a part of time seriesdecomposition result.

has_step_changes()

If true, step_changes is a part of time series decompositionresult.

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

class ClusterInfo(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

Information about a single cluster for clustering model.

centroid_id()

Centroid id.

cluster_radius()

Cluster radius, the average distance fromcentroid to each point assigned to the cluster.

cluster_size()

Cluster size, the total number of pointsassigned to the cluster.

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

class TrainingOptions(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

Options used in model training.

max_iterations()

The maximum number of iterations in training.Used only for iterative training algorithms.

loss_type()

Type of loss function used during trainingrun.

  • Type

    google.cloud.bigquery_v2.types.Model.LossType

learn_rate()

Learning rate in training. Used only foriterative training algorithms.

l1_regularization()

L1 regularization coefficient.

l2_regularization()

L2 regularization coefficient.

min_relative_progress()

When early_stop is true, stops training when accuracyimprovement is less than ‘min_relative_progress’. Used onlyfor iterative training algorithms.

warm_start()

Whether to train a model from the lastcheckpoint.

early_stop()

Whether to stop early when the loss doesn’t improvesignificantly any more (compared to min_relative_progress).Used only for iterative training algorithms.

input_label_columns()

Name of input label columns in training data.

  • Type

    Sequence[str]

data_split_method()

The data split type for training andevaluation, e.g. RANDOM.

  • Type

    google.cloud.bigquery_v2.types.Model.DataSplitMethod

data_split_eval_fraction()

The fraction of evaluation data over thewhole input data. The rest of data will be usedas training data. The format should be double.Accurate to two decimal places.Default value is 0.2.

data_split_column()

The column to split data with. This column won’t be used asa feature.

  1. When data_split_method is CUSTOM, the correspondingcolumn should be boolean. The rows with true value tagare eval data, and the false are training data.

  2. When data_split_method is SEQ, the firstDATA_SPLIT_EVAL_FRACTION rows (from smallest to largest)in the corresponding column are used as training data,and the rest are eval data. It respects the order inOrderable data types:https://cloud.google.com/bigquery/docs/reference/standard-sql/data-types#data-type-properties

learn_rate_strategy()

The strategy to determine learn rate for thecurrent iteration.

  • Type

    google.cloud.bigquery_v2.types.Model.LearnRateStrategy

initial_learn_rate()

Specifies the initial learning rate for theline search learn rate strategy.

label_class_weights()

Weights associated with each label class, forrebalancing the training data. Only applicablefor classification models.

user_column()

User column specified for matrixfactorization models.

item_column()

Item column specified for matrixfactorization models.

distance_type()

Distance type for clustering models.

  • Type

    google.cloud.bigquery_v2.types.Model.DistanceType

num_clusters()

Number of clusters for clustering models.

model_uri()

Google Cloud Storage URI from which the modelwas imported. Only applicable for importedmodels.

optimization_strategy()

Optimization strategy for training linearregression models.

  • Type

    google.cloud.bigquery_v2.types.Model.OptimizationStrategy

hidden_units()

Hidden units for dnn models.

  • Type

    Sequence[int]

batch_size()

Batch size for dnn models.

dropout()

Dropout probability for dnn models.

max_tree_depth()

Maximum depth of a tree for boosted treemodels.

subsample()

Subsample fraction of the training data togrow tree to prevent overfitting for boostedtree models.

min_split_loss()

Minimum split loss for boosted tree models.

num_factors()

Num factors specified for matrixfactorization models.

feedback_type()

Feedback type that specifies which algorithmto run for matrix factorization.

  • Type

    google.cloud.bigquery_v2.types.Model.FeedbackType

wals_alpha()

Hyperparameter for matrix factoration whenimplicit feedback type is specified.

kmeans_initialization_method()

The method used to initialize the centroidsfor kmeans algorithm.

  • Type

    google.cloud.bigquery_v2.types.Model.KmeansEnums.KmeansInitializationMethod

kmeans_initialization_column()

The column used to provide the initial centroids for kmeansalgorithm when kmeans_initialization_method is CUSTOM.

time_series_timestamp_column()

Column to be designated as time seriestimestamp for ARIMA model.

time_series_data_column()

Column to be designated as time series datafor ARIMA model.

auto_arima()

Whether to enable auto ARIMA or not.

non_seasonal_order()

A specification of the non-seasonal part ofthe ARIMA model: the three components (p, d, q)are the AR order, the degree of differencing,and the MA order.

  • Type

    google.cloud.bigquery_v2.types.Model.ArimaOrder

data_frequency()

The data frequency of a time series.

  • Type

    google.cloud.bigquery_v2.types.Model.DataFrequency

include_drift()

Include drift when fitting an ARIMA model.

holiday_region()

The geographical region based on which theholidays are considered in time series modeling.If a valid value is specified, then holidayeffects modeling is enabled.

  • Type

    google.cloud.bigquery_v2.types.Model.HolidayRegion

time_series_id_column()

The time series id column that was usedduring ARIMA model training.

time_series_id_columns()

The time series id columns that were usedduring ARIMA model training.

  • Type

    Sequence[str]

horizon()

The number of periods ahead that need to beforecasted.

preserve_input_structs()

Whether to preserve the input structs in output featurenames. Suppose there is a struct A with field b. When false(default), the output feature name is A_b. When true, theoutput feature name is A.b.

auto_arima_max_order()

The max value of non-seasonal p and q.

decompose_time_series()

If true, perform decompose time series andsave the results.

clean_spikes_and_dips()

If true, clean spikes and dips in the inputtime series.

adjust_step_changes()

If true, detect step changes and make dataadjustment in the input time series.

class LabelClassWeightsEntry(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

class google.cloud.bigquery_v2.types.ModelReference(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

Id path of a model.

project_id()

Required. The ID of the project containingthis model.

dataset_id()

Required. The ID of the dataset containingthis model.

model_id()

Required. The ID of the model. The ID must contain onlyletters (a-z, A-Z), numbers (0-9), or underscores (_). Themaximum length is 1,024 characters.

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

class google.cloud.bigquery_v2.types.PatchModelRequest(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

project_id()

Required. Project ID of the model to patch.

dataset_id()

Required. Dataset ID of the model to patch.

model_id()

Required. Model ID of the model to patch.

model()

Required. Patched model.Follows RFC5789 patch semantics. Missing fieldsare not updated. To clear a field, explicitlyset to default value.

  • Type

    google.cloud.bigquery_v2.types.Model

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

class google.cloud.bigquery_v2.types.StandardSqlDataType(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

The type of a variable, e.g., a function argument. Examples: INT64:{type_kind=”INT64”} ARRAY: {type_kind=”ARRAY”,array_element_type=”STRING”} STRUCT<x STRING, y ARRAY>:{type_kind=”STRUCT”, struct_type={fields=[ {name=”x”,type={type_kind=”STRING”}}, {name=”y”, type={type_kind=”ARRAY”,array_element_type=”DATE”}} ]}}

This message hasoneof fields (mutually exclusive fields).For each oneof, at most one member field can be set at the same time.Setting any member of the oneof automatically clears all othermembers.

type_kind()

Required. The top level type of this field.Can be any standard SQL data type (e.g.,“INT64”, “DATE”, “ARRAY”).

  • Type

    google.cloud.bigquery_v2.types.StandardSqlDataType.TypeKind

array_element_type()

The type of the array’s elements, if type_kind = “ARRAY”.

This field is a member ofoneofsub_type.

  • Type

    google.cloud.bigquery_v2.types.StandardSqlDataType

struct_type()

The fields of this struct, in order, if type_kind =“STRUCT”.

This field is a member ofoneofsub_type.

  • Type

    google.cloud.bigquery_v2.types.StandardSqlStructType

class TypeKind(value)

Bases:proto.enums.Enum

ARRAY( = 1 )

BIGNUMERIC( = 2 )

BOOL( = )

BYTES( = )

DATE( = 1 )

DATETIME( = 2 )

FLOAT64( = )

GEOGRAPHY( = 2 )

INT64( = )

INTERVAL( = 2 )

JSON( = 2 )

NUMERIC( = 2 )

STRING( = )

STRUCT( = 1 )

TIME( = 2 )

TIMESTAMP( = 1 )

TYPE_KIND_UNSPECIFIED( = )

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

class google.cloud.bigquery_v2.types.StandardSqlField(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

A field or a column.

name()

Optional. The name of this field. Can beabsent for struct fields.

type()

Optional. The type of this parameter. Absentif not explicitly specified (e.g., CREATEFUNCTION statement can omit the return type; inthis case the output parameter does not havethis “type” field).

  • Type

    google.cloud.bigquery_v2.types.StandardSqlDataType

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

class google.cloud.bigquery_v2.types.StandardSqlStructType(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

fields()

  • Type

    Sequence[google.cloud.bigquery_v2.types.StandardSqlField]

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

class google.cloud.bigquery_v2.types.StandardSqlTableType(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

A table type

columns()

The columns in this table type

  • Type

    Sequence[google.cloud.bigquery_v2.types.StandardSqlField]

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

class google.cloud.bigquery_v2.types.TableReference(mapping=None, *, ignore_unknown_fields=False, **kwargs)

Bases:proto.message.Message

project_id()

Required. The ID of the project containingthis table.

dataset_id()

Required. The ID of the dataset containingthis table.

table_id()

Required. The ID of the table. The ID must contain onlyletters (a-z, A-Z), numbers (0-9), or underscores (_). Themaximum length is 1,024 characters. Certain operations allowsuffixing of the table ID with a partition decorator, suchassample_table$20190123.

project_id_alternative()

The alternative field that will be used when ESF is not ableto translate the received data to the project_id field.

  • Type

    Sequence[str]

dataset_id_alternative()

The alternative field that will be used when ESF is not ableto translate the received data to the project_id field.

  • Type

    Sequence[str]

table_id_alternative()

The alternative field that will be used when ESF is not ableto translate the received data to the project_id field.

  • Type

    Sequence[str]

_delattr_(key)

Delete the value on the given field.

This is generally equivalent to setting a falsy value.

_eq_(other)

Return True if the messages are equal, False otherwise.

_ne_(other)

Return True if the messages are unequal, False otherwise.

_setattr_(key, value)

Set the value on the given field.

For well-known protocol buffer types which are marshalled, eitherthe protocol buffer object or the Python equivalent is accepted.

Except as otherwise noted, the content of this page is licensed under theCreative Commons Attribution 4.0 License, and code samples are licensed under theApache 2.0 License. For details, see theGoogle Developers Site Policies. Java is a registered trademark of Oracle and/or its affiliates.

Last updated 2025-12-16 UTC.