BigQuery - Package cloud.google.com/go/bigquery (v1.57.1)

Package bigquery provides a client for the BigQuery service.

The following assumes a basic familiarity with BigQuery concepts. See https://cloud.google.com/bigquery/docs.

See https://godoc.org/cloud.google.com/go for authentication, timeouts, connection pooling and similar aspects of this package.

Creating a Client

To start working with this package, create a client:

ctx := context.Background()
client, err := bigquery.NewClient(ctx, projectID)
if err != nil {
    // TODO: Handle error.
}

Querying

To query existing tables, create a Query and call its Read method:

q := client.Query(`
    SELECT year, SUM(number) as num
    FROM ` + "`bigquery-public-data.usa_names.usa_1910_2013`" + `
    WHERE name = "William"
    GROUP BY year
    ORDER BY year
`)
it, err := q.Read(ctx)
if err != nil {
    // TODO: Handle error.
}

Then iterate through the resulting rows. You can store a row using anything that implements the ValueLoader interface, or with a slice or map of bigquery.Value. A slice is simplest:

for {
    var values []bigquery.Value
    err := it.Next(&values)
    if err == iterator.Done {
        break
    }
    if err != nil {
        // TODO: Handle error.
    }
    fmt.Println(values)
}

You can also use a struct whose exported fields match the query:

type Count struct {
    Year int
    Num  int
}
for {
    var c Count
    err := it.Next(&c)
    if err == iterator.Done {
        break
    }
    if err != nil {
        // TODO: Handle error.
    }
    fmt.Println(c)
}

You can also start the query running and get the results later. Create the query as above, but call Run instead of Read. This returns a Job, which represents an asynchronous operation.

job, err := q.Run(ctx)
if err != nil {
    // TODO: Handle error.
}

Get the job's ID, a printable string. You can save this string to retrieve the results at a later time, even in another process.

jobID := job.ID()
fmt.Printf("The job ID is %s\n", jobID)

To retrieve the job's results from the ID, first look up the Job:

job, err = client.JobFromID(ctx, jobID)
if err != nil {
    // TODO: Handle error.
}

Use the Job.Read method to obtain an iterator, and loop over the rows. Calling Query.Read is preferred for queries with a relatively small result set, as it will call BigQuery jobs.query API for a optimized query path. If the query doesn't meet that criteria, the method will just combine Query.Run and Job.Read.

it, err = job.Read(ctx)
if err != nil {
    // TODO: Handle error.
}
// Proceed with iteration as above.

Datasets and Tables

You can refer to datasets in the client's project with the Dataset method, and in other projects with the DatasetInProject method:

myDataset := client.Dataset("my_dataset")
yourDataset := client.DatasetInProject("your-project-id", "your_dataset")

These methods create references to datasets, not the datasets themselves. You can have a dataset reference even if the dataset doesn't exist yet. Use Dataset.Create to create a dataset from a reference:

if err := myDataset.Create(ctx, nil); err != nil {
    // TODO: Handle error.
}

You can refer to tables with Dataset.Table. Like bigquery.Dataset, bigquery.Table is a reference to an object in BigQuery that may or may not exist.

table := myDataset.Table("my_table")

You can create, delete and update the metadata of tables with methods on Table. For instance, you could create a temporary table with:

err = myDataset.Table("temp").Create(ctx, &bigquery.TableMetadata{
    ExpirationTime: time.Now().Add(1*time.Hour)})
if err != nil {
    // TODO: Handle error.
}

We'll see how to create a table with a schema in the next section.

Schemas

There are two ways to construct schemas with this package. You can build a schema by hand, like so:

schema1 := bigquery.Schema{
    {Name: "Name", Required: true, Type: bigquery.StringFieldType},
    {Name: "Grades", Repeated: true, Type: bigquery.IntegerFieldType},
    {Name: "Optional", Required: false, Type: bigquery.IntegerFieldType},
}

Or you can infer the schema from a struct:

type student struct {
    Name   string
    Grades []int
    Optional bigquery.NullInt64
}
schema2, err := bigquery.InferSchema(student{})
if err != nil {
    // TODO: Handle error.
}
// schema1 and schema2 are identical.

Struct inference supports tags like those of the encoding/json package, so you can change names, ignore fields, or mark a field as nullable (non-required). Fields declared as one of the Null types (NullInt64, NullFloat64, NullString, NullBool, NullTimestamp, NullDate, NullTime, NullDateTime, and NullGeography) are automatically inferred as nullable, so the "nullable" tag is only needed for []byte, *big.Rat and pointer-to-struct fields.

type student2 struct {
    Name     string `bigquery:"full_name"`
    Grades   []int
    Secret   string `bigquery:"-"`
    Optional []byte `bigquery:",nullable"`
}
schema3, err := bigquery.InferSchema(student2{})
if err != nil {
    // TODO: Handle error.
}
// schema3 has required fields "full_name" and "Grade", and nullable BYTES field "Optional".

Having constructed a schema, you can create a table with it like so:

if err := table.Create(ctx, &bigquery.TableMetadata{Schema: schema1}); err != nil {
    // TODO: Handle error.
}

Copying

You can copy one or more tables to another table. Begin by constructing a Copier describing the copy. Then set any desired copy options, and finally call Run to get a Job:

copier := myDataset.Table("dest").CopierFrom(myDataset.Table("src"))
copier.WriteDisposition = bigquery.WriteTruncate
job, err = copier.Run(ctx)
if err != nil {
    // TODO: Handle error.
}

You can chain the call to Run if you don't want to set options:

job, err = myDataset.Table("dest").CopierFrom(myDataset.Table("src")).Run(ctx)
if err != nil {
    // TODO: Handle error.
}

You can wait for your job to complete:

status, err := job.Wait(ctx)
if err != nil {
    // TODO: Handle error.
}

Job.Wait polls with exponential backoff. You can also poll yourself, if you wish:

for {
    status, err := job.Status(ctx)
    if err != nil {
        // TODO: Handle error.
    }
    if status.Done() {
        if status.Err() != nil {
            log.Fatalf("Job failed with error %v", status.Err())
        }
        break
    }
    time.Sleep(pollInterval)
}

Loading and Uploading

There are two ways to populate a table with this package: load the data from a Google Cloud Storage object, or upload rows directly from your program.

For loading, first create a GCSReference, configuring it if desired. Then make a Loader, optionally configure it as well, and call its Run method.

gcsRef := bigquery.NewGCSReference("gs://my-bucket/my-object")
gcsRef.AllowJaggedRows = true
loader := myDataset.Table("dest").LoaderFrom(gcsRef)
loader.CreateDisposition = bigquery.CreateNever
job, err = loader.Run(ctx)
// Poll the job for completion if desired, as above.

To upload, first define a type that implements the ValueSaver interface, which has a single method named Save. Then create an Inserter, and call its Put method with a slice of values.

u := table.Inserter()
// Item implements the ValueSaver interface.
items := []*Item{
    {Name: "n1", Size: 32.6, Count: 7},
    {Name: "n2", Size: 4, Count: 2},
    {Name: "n3", Size: 101.5, Count: 1},
}
if err := u.Put(ctx, items); err != nil {
    // TODO: Handle error.
}

You can also upload a struct that doesn't implement ValueSaver. Use the StructSaver type to specify the schema and insert ID by hand, or just supply the struct or struct pointer directly and the schema will be inferred:

type Item2 struct {
    Name  string
    Size  float64
    Count int
}
// Item implements the ValueSaver interface.
items2 := []*Item2{
    {Name: "n1", Size: 32.6, Count: 7},
    {Name: "n2", Size: 4, Count: 2},
    {Name: "n3", Size: 101.5, Count: 1},
}
if err := u.Put(ctx, items2); err != nil {
    // TODO: Handle error.
}

BigQuery allows for higher throughput when omitting insertion IDs. To enable this, specify the sentinel NoDedupeID value for the insertion ID when implementing a ValueSaver.

Extracting

If you've been following so far, extracting data from a BigQuery table into a Google Cloud Storage object will feel familiar. First create an Extractor, then optionally configure it, and lastly call its Run method.

extractor := table.ExtractorTo(gcsRef)
extractor.DisableHeader = true
job, err = extractor.Run(ctx)
// Poll the job for completion if desired, as above.

Errors

Errors returned by this client are often of the type googleapi.Error: https://godoc.org/google.golang.org/api/googleapi#Error

These errors can be introspected for more information by using xerrors.As with the richer *googleapi.Error type. For example:

   var e *googleapi.Error
    if ok := xerrors.As(err, &e); ok {
          if e.Code == 409 { ... }
    }

In some cases, your client may received unstructured googleapi.Error error responses. In such cases, it is likely that you have exceeded BigQuery request limits, documented at: https://cloud.google.com/bigquery/quotas

Constants

LogicalStorageBillingModel, PhysicalStorageBillingModel

const (
	// LogicalStorageBillingModel indicates billing for logical bytes.
	LogicalStorageBillingModel = ""

	// PhysicalStorageBillingModel indicates billing for physical bytes.
	PhysicalStorageBillingModel = "PHYSICAL"
)

ScalarFunctionRoutine, ProcedureRoutine, TableValuedFunctionRoutine

const (
	// ScalarFunctionRoutine scalar function routine type
	ScalarFunctionRoutine = "SCALAR_FUNCTION"
	// ProcedureRoutine procedure routine type
	ProcedureRoutine = "PROCEDURE"
	// TableValuedFunctionRoutine routine type for table valued functions
	TableValuedFunctionRoutine = "TABLE_VALUED_FUNCTION"
)

NumericPrecisionDigits, NumericScaleDigits, BigNumericPrecisionDigits, BigNumericScaleDigits

const (
	// NumericPrecisionDigits is the maximum number of digits in a NUMERIC value.
	NumericPrecisionDigits = 38

	// NumericScaleDigits is the maximum number of digits after the decimal point in a NUMERIC value.
	NumericScaleDigits = 9

	// BigNumericPrecisionDigits is the maximum number of full digits in a BIGNUMERIC value.
	BigNumericPrecisionDigits = 76

	// BigNumericScaleDigits is the maximum number of full digits in a BIGNUMERIC value.
	BigNumericScaleDigits = 38
)

DetectProjectID

const DetectProjectID = "*detect-project-id*"

DetectProjectID is a sentinel value that instructs NewClient to detect the project ID. It is given in place of the projectID argument. NewClient will use the project ID from the given credentials or the default credentials (https://developers.google.com/accounts/docs/application-default-credentials) if no credentials were provided. When providing credentials, not all options will allow NewClient to extract the project ID. Specifically a JWT does not have the project ID encoded.

NoDedupeID

const NoDedupeID = "NoDedupeID"

NoDedupeID indicates a streaming insert row wants to opt out of best-effort deduplication. It is EXPERIMENTAL and subject to change or removal without notice.

Scope

const (
	// Scope is the Oauth2 scope for the service.
	// For relevant BigQuery scopes, see:
	// https://developers.google.com/identity/protocols/googlescopes#bigqueryv2
	Scope = "https://www.googleapis.com/auth/bigquery"
)

Variables

NeverExpire

var NeverExpire = time.Time{}.Add(-1)

NeverExpire is a sentinel value used to remove a table'e expiration time.

Functions

func BigNumericString 

func BigNumericString(r *big.Rat) string

BigNumericString returns a string representing a *big.Rat in a format compatible with BigQuery SQL. It returns a floating point literal with 38 digits after the decimal point.

func CivilDateTimeString 

func CivilDateTimeString(dt civil.DateTime) string

CivilDateTimeString returns a string representing a civil.DateTime in a format compatible with BigQuery SQL. It separate the date and time with a space, and formats the time with CivilTimeString.

Use CivilDateTimeString when using civil.DateTime in DML, for example in INSERT statements.

func CivilTimeString 

func CivilTimeString(t civil.Time) string

CivilTimeString returns a string representing a civil.Time in a format compatible with BigQuery SQL. It rounds the time to the nearest microsecond and returns a string with six digits of sub-second precision.

Use CivilTimeString when using civil.Time in DML, for example in INSERT statements.

func IntervalString 

func IntervalString(iv *IntervalValue) string

IntervalString returns a string representing an *IntervalValue in a format compatible with BigQuery SQL. It returns an interval literal in canonical format.

func NewArrowIteratorReader 

func NewArrowIteratorReader(it ArrowIterator) io.Reader

NewArrowIteratorReader allows to consume an ArrowIterator as an io.Reader. Experimental: this interface is experimental and may be modified or removed in future versions, regardless of any other documented package stability guarantees.

func NumericString 

func NumericString(r *big.Rat) string

NumericString returns a string representing a *big.Rat in a format compatible with BigQuery SQL. It returns a floating-point literal with 9 digits after the decimal point.

func Seed 

func Seed(s int64)

Seed seeds this package's random number generator, used for generating job and insert IDs. Use Seed to obtain repeatable, deterministic behavior from bigquery clients. Seed should be called before any clients are created.

AccessEntry

type AccessEntry struct {
	Role       AccessRole          // The role of the entity
	EntityType EntityType          // The type of entity
	Entity     string              // The entity (individual or group) granted access
	View       *Table              // The view granted access (EntityType must be ViewEntity)
	Routine    *Routine            // The routine granted access (only UDF currently supported)
	Dataset    *DatasetAccessEntry // The resources within a dataset granted access.
}

An AccessEntry describes the permissions that an entity has on a dataset.

AccessRole

type AccessRole string

AccessRole is the level of access to grant to a dataset.

OwnerRole, ReaderRole, WriterRole

const (
	// OwnerRole is the OWNER AccessRole.
	OwnerRole AccessRole = "OWNER"
	// ReaderRole is the READER AccessRole.
	ReaderRole AccessRole = "READER"
	// WriterRole is the WRITER AccessRole.
	WriterRole AccessRole = "WRITER"
)

ArrowIterator

type ArrowIterator interface {
	Next() (*ArrowRecordBatch, error)
	Schema() Schema
	SerializedArrowSchema() []byte
}

ArrowIterator represents a way to iterate through a stream of arrow records. Experimental: this interface is experimental and may be modified or removed in future versions, regardless of any other documented package stability guarantees.

ArrowRecordBatch

type ArrowRecordBatch struct {

	// Serialized Arrow Record Batch.
	Data []byte
	// Serialized Arrow Schema.
	Schema []byte
	// Source partition ID. In the Storage API world, it represents the ReadStream.
	PartitionID string
	// contains filtered or unexported fields
}

ArrowRecordBatch represents an Arrow RecordBatch with the source PartitionID

func (*ArrowRecordBatch) Read 

func (r *ArrowRecordBatch) Read(p []byte) (int, error)

Read makes ArrowRecordBatch implements io.Reader

AvroOptions

type AvroOptions struct {
	// UseAvroLogicalTypes indicates whether to interpret logical types as the
	// corresponding BigQuery data type (for example, TIMESTAMP), instead of using
	// the raw type (for example, INTEGER).
	UseAvroLogicalTypes bool
}

AvroOptions are additional options for Avro external data data sources.

BIEngineReason

type BIEngineReason struct {
	// High-Level BI engine reason for partial or disabled acceleration.
	Code string

	// Human-readable reason for partial or disabled acceleration.
	Message string
}

BIEngineReason contains more detailed information about why a query wasn't fully accelerated.

BIEngineStatistics

type BIEngineStatistics struct {
	// Specifies which mode of BI Engine acceleration was performed.
	BIEngineMode string

	// In case of DISABLED or PARTIAL BIEngineMode, these
	// contain the explanatory reasons as to why BI Engine could not
	// accelerate. In case the full query was accelerated, this field is not
	// populated.
	BIEngineReasons []*BIEngineReason
}

BIEngineStatistics contains query statistics specific to the use of BI Engine.

BigtableColumn

type BigtableColumn struct {
	// Qualifier of the column. Columns in the parent column family that have this
	// exact qualifier are exposed as . field. The column field name is the
	// same as the column qualifier.
	Qualifier string

	// If the qualifier is not a valid BigQuery field identifier i.e. does not match
	// [a-zA-Z][a-zA-Z0-9_]*, a valid identifier must be provided as the column field
	// name and is used as field name in queries.
	FieldName string

	// If true, only the latest version of values are exposed for this column.
	// See BigtableColumnFamily.OnlyReadLatest.
	OnlyReadLatest bool

	// The encoding of the values when the type is not STRING.
	// See BigtableColumnFamily.Encoding
	Encoding string

	// The type to convert the value in cells of this column.
	// See BigtableColumnFamily.Type
	Type string
}

BigtableColumn describes how BigQuery should access a Bigtable column.

BigtableColumnFamily

type BigtableColumnFamily struct {
	// Identifier of the column family.
	FamilyID string

	// Lists of columns that should be exposed as individual fields as opposed to a
	// list of (column name, value) pairs. All columns whose qualifier matches a
	// qualifier in this list can be accessed as .. Other columns can be accessed as
	// a list through .Column field.
	Columns []*BigtableColumn

	// The encoding of the values when the type is not STRING. Acceptable encoding values are:
	// - TEXT - indicates values are alphanumeric text strings.
	// - BINARY - indicates values are encoded using HBase Bytes.toBytes family of functions.
	// This can be overridden for a specific column by listing that column in 'columns' and
	// specifying an encoding for it.
	Encoding string

	// If true, only the latest version of values are exposed for all columns in this
	// column family. This can be overridden for a specific column by listing that
	// column in 'columns' and specifying a different setting for that column.
	OnlyReadLatest bool

	// The type to convert the value in cells of this
	// column family. The values are expected to be encoded using HBase
	// Bytes.toBytes function when using the BINARY encoding value.
	// Following BigQuery types are allowed (case-sensitive):
	// BYTES STRING INTEGER FLOAT BOOLEAN.
	// The default type is BYTES. This can be overridden for a specific column by
	// listing that column in 'columns' and specifying a type for it.
	Type string
}

BigtableColumnFamily describes how BigQuery should access a Bigtable column family.

BigtableOptions

type BigtableOptions struct {
	// A list of column families to expose in the table schema along with their
	// types. If omitted, all column families are present in the table schema and
	// their values are read as BYTES.
	ColumnFamilies []*BigtableColumnFamily

	// If true, then the column families that are not specified in columnFamilies
	// list are not exposed in the table schema. Otherwise, they are read with BYTES
	// type values. The default is false.
	IgnoreUnspecifiedColumnFamilies bool

	// If true, then the rowkey column families will be read and converted to string.
	// Otherwise they are read with BYTES type values and users need to manually cast
	// them with CAST if necessary. The default is false.
	ReadRowkeyAsString bool
}

BigtableOptions are additional options for Bigtable external data sources.

CSVOptions

type CSVOptions struct {
	// AllowJaggedRows causes missing trailing optional columns to be tolerated
	// when reading CSV data. Missing values are treated as nulls.
	AllowJaggedRows bool

	// AllowQuotedNewlines sets whether quoted data sections containing
	// newlines are allowed when reading CSV data.
	AllowQuotedNewlines bool

	// Encoding is the character encoding of data to be read.
	Encoding Encoding

	// FieldDelimiter is the separator for fields in a CSV file, used when
	// reading or exporting data. The default is ",".
	FieldDelimiter string

	// Quote is the value used to quote data sections in a CSV file. The
	// default quotation character is the double quote ("), which is used if
	// both Quote and ForceZeroQuote are unset.
	// To specify that no character should be interpreted as a quotation
	// character, set ForceZeroQuote to true.
	// Only used when reading data.
	Quote          string
	ForceZeroQuote bool

	// The number of rows at the top of a CSV file that BigQuery will skip when
	// reading data.
	SkipLeadingRows int64

	// An optional custom string that will represent a NULL
	// value in CSV import data.
	NullMarker string

	// Preserves the embedded ASCII control characters (the first 32 characters in the ASCII-table,
	// from '\\x00' to '\\x1F') when loading from CSV. Only applicable to CSV, ignored for other formats.
	PreserveASCIIControlCharacters bool
}

CSVOptions are additional options for CSV external data sources.

Client

type Client struct {
	// Location, if set, will be used as the default location for all subsequent
	// dataset creation and job operations. A location specified directly in one of
	// those operations will override this value.
	Location string
	// contains filtered or unexported fields
}

Client may be used to perform BigQuery operations.

func NewClient 

func NewClient(ctx context.Context, projectID string, opts ...option.ClientOption) (*Client, error)

NewClient constructs a new Client which can perform BigQuery operations. Operations performed via the client are billed to the specified GCP project.

If the project ID is set to DetectProjectID, NewClient will attempt to detect the project ID from credentials.

This client supports enabling query-related preview features via environmental variables. By setting the environment variable QUERY_PREVIEW_ENABLED to the string "TRUE", the client will enable preview features, though behavior may still be controlled via the bigquery service as well. Currently, the feature(s) in scope include: stateless queries (query execution without corresponding job metadata).

Example

package main

import (
	"context"

	"cloud.google.com/go/bigquery"
)

func main() {
	ctx := context.Background()
	client, err := bigquery.NewClient(ctx, "project-id")
	if err != nil {
		// TODO: Handle error.
	}
	_ = client // TODO: Use client.
}

func (*Client) Close 

func (c *Client) Close() error

Close closes any resources held by the client. Close should be called when the client is no longer needed. It need not be called at program exit.

func (*Client) Dataset 

func (c *Client) Dataset(id string) *Dataset

Dataset creates a handle to a BigQuery dataset in the client's project.

Example

package main

import (
	"context"
	"fmt"

	"cloud.google.com/go/bigquery"
)

func main() {
	ctx := context.Background()
	client, err := bigquery.NewClient(ctx, "project-id")
	if err != nil {
		// TODO: Handle error.
	}
	ds := client.Dataset("my_dataset")
	fmt.Println(ds)
}

func (*Client) DatasetInProject 

func (c *Client) DatasetInProject(projectID, datasetID string) *Dataset

DatasetInProject creates a handle to a BigQuery dataset in the specified project.

Example

package main

import (
	"context"
	"fmt"

	"cloud.google.com/go/bigquery"
)

func main() {
	ctx := context.Background()
	client, err := bigquery.NewClient(ctx, "project-id")
	if err != nil {
		// TODO: Handle error.
	}
	ds := client.DatasetInProject("their-project-id", "their-dataset")
	fmt.Println(ds)
}

func (*Client) Datasets 

func (c *Client) Datasets(ctx context.Context) *DatasetIterator

Datasets returns an iterator over the datasets in a project. The Client's project is used by default, but that can be changed by setting ProjectID on the returned iterator before calling Next.

Example

package main

import (
	"context"

	"cloud.google.com/go/bigquery"
)

func main() {
	ctx := context.Background()
	client, err := bigquery.NewClient(ctx, "project-id")
	if err != nil {
		// TODO: Handle error.
	}
	it := client.Datasets(ctx)
	_ = it // TODO: iterate using Next or iterator.Pager.
}

func (*Client) DatasetsInProject (deprecated)

func (c *Client) DatasetsInProject(ctx context.Context, projectID string) *DatasetIterator

DatasetsInProject returns an iterator over the datasets in the provided project.

Deprecated: call Client.Datasets, then set ProjectID on the returned iterator.

Example

package main

import (
	"context"

	"cloud.google.com/go/bigquery"
)

func main() {
	ctx := context.Background()
	client, err := bigquery.NewClient(ctx, "project-id")
	if err != nil {
		// TODO: Handle error.
	}
	it := client.DatasetsInProject(ctx, "their-project-id")
	_ = it // TODO: iterate using Next or iterator.Pager.
}

func (*Client) EnableStorageReadClient 

func (c *Client) EnableStorageReadClient(ctx context.Context, opts ...option.ClientOption) error

EnableStorageReadClient sets up Storage API connection to be used when fetching large datasets from tables, jobs or queries. Currently out of pagination methods like PageInfo().Token and RowIterator.StartIndex are not supported when the Storage API is enabled. Calling this method twice will return an error.

func (*Client) JobFromID 

func (c *Client) JobFromID(ctx context.Context, id string) (*Job, error)

JobFromID creates a Job which refers to an existing BigQuery job. The job need not have been created by this package. For example, the job may have been created in the BigQuery console.

For jobs whose location is other than "US" or "EU", set Client.Location or use JobFromIDLocation.

Example

package main

import (
	"context"
	"fmt"

	"cloud.google.com/go/bigquery"
)

func getJobID() string { return "" }

func main() {
	ctx := context.Background()
	client, err := bigquery.NewClient(ctx, "project-id")
	if err != nil {
		// TODO: Handle error.
	}
	jobID := getJobID() // Get a job ID using Job.ID, the console or elsewhere.
	job, err := client.JobFromID(ctx, jobID)
	if err != nil {
		// TODO: Handle error.
	}
	fmt.Println(job.LastStatus()) // Display the job's status.
}

func (*Client) JobFromIDLocation 

func (c *Client) JobFromIDLocation(ctx context.Context, id, location string) (j *Job, err error)

JobFromIDLocation creates a Job which refers to an existing BigQuery job. The job need not have been created by this package (for example, it may have been created in the BigQuery console), but it must exist in the specified location.

func (*Client) JobFromProject 

func (c *Client) JobFromProject(ctx context.Context, projectID, jobID, location string) (j *Job, err error)

JobFromProject creates a Job which refers to an existing BigQuery job. The job need not have been created by this package, nor does it need to reside within the same project or location as the instantiated client.

func (*Client) Jobs 

func (c *Client) Jobs(ctx context.Context) *JobIterator

Jobs lists jobs within a project.

Example

package main

import (
	"context"

	"cloud.google.com/go/bigquery"
)

func main() {
	ctx := context.Background()
	client, err := bigquery.NewClient(ctx, "project-id")
	if err != nil {
		// TODO: Handle error.
	}
	it := client.Jobs(ctx)
	it.State = bigquery.Running // list only running jobs.
	_ = it                      // TODO: iterate using Next or iterator.Pager.
}

func (*Client) Project 

func (c *Client) Project() string

Project returns the project ID or number for this instance of the client, which may have either been explicitly specified or autodetected.

func (*Client) Query 

func (c *Client) Query(q string) *Query

Query creates a query with string q. The returned Query may optionally be further configured before its Run method is called.

Examples

package main

import (
	"context"

	"cloud.google.com/go/bigquery"
)

func main() {
	ctx := context.Background()
	client, err := bigquery.NewClient(ctx, "project-id")
	if err != nil {
		// TODO: Handle error.
	}
	q := client.Query("select name, num from t1")
	q.DefaultProjectID = "project-id"
	// TODO: set other options on the Query.
	// TODO: Call Query.Run or Query.Read.
}
encryptionKey
package main

import (
	"context"

	"cloud.google.com/go/bigquery"
)

func main() {
	ctx := context.Background()
	client, err := bigquery.NewClient(ctx, "project-id")
	if err != nil {
		// TODO: Handle error.
	}
	q := client.Query("select name, num from t1")
	// TODO: Replace this key with a key you have created in Cloud KMS.
	keyName := "projects/P/locations/L/keyRings/R/cryptoKeys/K"
	q.DestinationEncryptionConfig = &bigquery.EncryptionConfig{KMSKeyName: keyName}
	// TODO: set other options on the Query.
	// TODO: Call Query.Run or Query.Read.
}
parameters
package main

import (
	"context"

	"cloud.google.com/go/bigquery"
)

func main() {
	ctx := context.Background()
	client, err := bigquery.NewClient(ctx, "project-id")
	if err != nil {
		// TODO: Handle error.
	}
	q := client.Query("select num from t1 where name = @user")
	q.Parameters = []bigquery.QueryParameter{
		{Name: "user", Value: "Elizabeth"},
	}
	// TODO: set other options on the Query.
	// TODO: Call Query.Run or Query.Read.
}

CloneDefinition

type CloneDefinition struct {

	// BaseTableReference describes the ID of the table that this clone
	// came from.
	BaseTableReference *Table

	// CloneTime indicates when the base table was cloned.
	CloneTime time.Time
}

CloneDefinition provides metadata related to the origin of a clone.

Clustering

type Clustering struct {
	Fields []string
}

Clustering governs the organization of data within a managed table. For more information, see https://cloud.google.com/bigquery/docs/clustered-tables

ColumnReference

type ColumnReference struct {
	// ReferencingColumn is the column in the current table that composes the foreign key.
	ReferencingColumn string
	// ReferencedColumn is the column in the primary key of the foreign table that
	// is referenced by the ReferencingColumn.
	ReferencedColumn string
}

ColumnReference represents the pair of the foreign key column and primary key column.

Compression

type Compression string

Compression is the type of compression to apply when writing data to Google Cloud Storage.

None, Gzip, Deflate, Snappy

const (
	// None specifies no compression.
	None Compression = "NONE"
	// Gzip specifies gzip compression.
	Gzip Compression = "GZIP"
	// Deflate specifies DEFLATE compression for Avro files.
	Deflate Compression = "DEFLATE"
	// Snappy specifies SNAPPY compression for Avro files.
	Snappy Compression = "SNAPPY"
)

ConnectionProperty

type ConnectionProperty struct {
	// Name of the connection property to set.
	Key string
	// Value of the connection property.
	Value string
}

ConnectionProperty represents a single key and value pair that can be sent alongside a query request or load job.

Copier

type Copier struct {
	JobIDConfig
	CopyConfig
	// contains filtered or unexported fields
}

A Copier copies data into a BigQuery table from one or more BigQuery tables.

func (*Copier) Run 

func (c *Copier) Run(ctx context.Context) (*Job, error)

Run initiates a copy job.

CopyConfig

type CopyConfig struct {
	// Srcs are the tables from which data will be copied.
	Srcs []*Table

	// Dst is the table into which the data will be copied.
	Dst *Table

	// CreateDisposition specifies the circumstances under which the destination table will be created.
	// The default is CreateIfNeeded.
	CreateDisposition TableCreateDisposition

	// WriteDisposition specifies how existing data in the destination table is treated.
	// The default is WriteEmpty.
	WriteDisposition TableWriteDisposition

	// The labels associated with this job.
	Labels map[string]string

	// Custom encryption configuration (e.g., Cloud KMS keys).
	DestinationEncryptionConfig *EncryptionConfig

	// One of the supported operation types when executing a Table Copy jobs.  By default this
	// copies tables, but can also be set to perform snapshot or restore operations.
	OperationType TableCopyOperationType

	// Sets a best-effort deadline on a specific job.  If job execution exceeds this
	// timeout, BigQuery may attempt to cancel this work automatically.
	//
	// This deadline cannot be adjusted or removed once the job is created.  Consider
	// using Job.Cancel in situations where you need more dynamic behavior.
	//
	// Experimental: this option is experimental and may be modified or removed in future versions,
	// regardless of any other documented package stability guarantees.
	JobTimeout time.Duration
}

CopyConfig holds the configuration for a copy job.

DMLStatistics

type DMLStatistics struct {
	// Rows added by the statement.
	InsertedRowCount int64
	// Rows removed by the statement.
	DeletedRowCount int64
	// Rows changed by the statement.
	UpdatedRowCount int64
}

DMLStatistics contains counts of row mutations triggered by a DML query statement.

DataFormat

type DataFormat string

DataFormat describes the format of BigQuery table data.

CSV, Avro, JSON, DatastoreBackup, GoogleSheets, Bigtable, Parquet, ORC, TFSavedModel, XGBoostBooster, Iceberg

const (
	CSV             DataFormat = "CSV"
	Avro            DataFormat = "AVRO"
	JSON            DataFormat = "NEWLINE_DELIMITED_JSON"
	DatastoreBackup DataFormat = "DATASTORE_BACKUP"
	GoogleSheets    DataFormat = "GOOGLE_SHEETS"
	Bigtable        DataFormat = "BIGTABLE"
	Parquet         DataFormat = "PARQUET"
	ORC             DataFormat = "ORC"
	// For BQ ML Models, TensorFlow Saved Model format.
	TFSavedModel DataFormat = "ML_TF_SAVED_MODEL"
	// For BQ ML Models, xgBoost Booster format.
	XGBoostBooster DataFormat = "ML_XGBOOST_BOOSTER"
	Iceberg        DataFormat = "ICEBERG"
)

Constants describing the format of BigQuery table data.

Dataset

type Dataset struct {
	ProjectID string
	DatasetID string
	// contains filtered or unexported fields
}

Dataset is a reference to a BigQuery dataset.

func (*Dataset) Create 

func (d *Dataset) Create(ctx context.Context, md *DatasetMetadata) (err error)

Create creates a dataset in the BigQuery service. An error will be returned if the dataset already exists. Pass in a DatasetMetadata value to configure the dataset.

Example

package main

import (
	"context"

	"cloud.google.com/go/bigquery"
)

func main() {
	ctx := context.Background()
	client, err := bigquery.NewClient(ctx, "project-id")
	if err != nil {
		// TODO: Handle error.
	}
	ds := client.Dataset("my_dataset")
	if err := ds.Create(ctx, &bigquery.DatasetMetadata{Location: "EU"}); err != nil {
		// TODO: Handle error.
	}
}

func (*Dataset) Delete 

func (d *Dataset) Delete(ctx context.Context) (err error)

Delete deletes the dataset. Delete will fail if the dataset is not empty.

Example

package main

import (
	"context"

	"cloud.google.com/go/bigquery"
)

func main() {
	ctx := context.Background()
	client, err := bigquery.NewClient(ctx, "project-id")
	if err != nil {
		// TODO: Handle error.
	}
	if err := client.Dataset("my_dataset").Delete(ctx); err != nil {
		// TODO: Handle error.
	}
}

func (*Dataset) DeleteWithContents 

func (d *Dataset) DeleteWithContents(ctx context.Context) (err error)

DeleteWithContents deletes the dataset, as well as contained resources.

func (*Dataset) Identifier 

func (d *Dataset) Identifier(f IdentifierFormat) (string, error)

Identifier returns the ID of the dataset in the requested format.

For Standard SQL format, the identifier will be quoted if the ProjectID contains dash (-) characters.

func (*Dataset) Metadata 

func (d *Dataset) Metadata(ctx context.Context) (md *DatasetMetadata, err error)

Metadata fetches the metadata for the dataset.

Example

package main

import (
	"context"
	"fmt"

	"cloud.google.com/go/bigquery"
)

func main() {
	ctx := context.Background()
	client, err := bigquery.NewClient(ctx, "project-id")
	if err != nil {
		// TODO: Handle error.
	}
	md, err := client.Dataset("my_dataset").Metadata(ctx)
	if err != nil {
		// TODO: Handle error.
	}
	fmt.Println(md)
}

func (*Dataset) Model 

func (d *Dataset) Model(modelID string) *Model

Model creates a handle to a BigQuery model in the dataset. To determine if a model exists, call Model.Metadata. If the model does not already exist, you can create it via execution of a CREATE MODEL query.

func (*Dataset) Models 

func (d *Dataset) Models(ctx context.Context) *ModelIterator

Models returns an iterator over the models in the Dataset.

func (*Dataset) Routine 

func (d *Dataset) Routine(routineID string) *Routine

Routine creates a handle to a BigQuery routine in the dataset. To determine if a routine exists, call Routine.Metadata.

func (*Dataset) Routines 

func (d *Dataset) Routines(ctx context.Context) *RoutineIterator

Routines returns an iterator over the routines in the Dataset.

func (*Dataset) Table 

func (d *Dataset) Table(tableID string) *Table

Table creates a handle to a BigQuery table in the dataset. To determine if a table exists, call Table.Metadata. If the table does not already exist, use Table.Create to create it.

Example

package main

import (
	"context"
	"fmt"

	"cloud.google.com/go/bigquery"
)

func main() {
	ctx := context.Background()
	client, err := bigquery.NewClient(ctx, "project-id")
	if err != nil {
		// TODO: Handle error.
	}
	// Table creates a reference to the table. It does not create the actual
	// table in BigQuery; to do so, use Table.Create.
	t := client.Dataset("my_dataset").Table("my_table")
	fmt.Println(t)
}

func (*Dataset) Tables 

func (d *Dataset) Tables(ctx context.Context) *TableIterator

Tables returns an iterator over the tables in the Dataset.

Example

package main

import (
	"context"

	"cloud.google.com/go/bigquery"
)

func main() {
	ctx := context.Background()
	client, err := bigquery.NewClient(ctx, "project-id")
	if err != nil {
		// TODO: Handle error.
	}
	it := client.Dataset("my_dataset").Tables(ctx)
	_ = it // TODO: iterate using Next or iterator.Pager.
}

func (*Dataset) Update 

func (d *Dataset) Update(ctx context.Context, dm DatasetMetadataToUpdate, etag string) (md *DatasetMetadata, err error)

Update modifies specific Dataset metadata fields. To perform a read-modify-write that protects against intervening reads, set the etag argument to the DatasetMetadata.ETag field from the read. Pass the empty string for etag for a "blind write" that will always succeed.

Examples

blindWrite
package main

import (
	"context"
	"fmt"

	"cloud.google.com/go/bigquery"
)

func main() {
	ctx := context.Background()
	client, err := bigquery.NewClient(ctx, "project-id")
	if err != nil {
		// TODO: Handle error.
	}
	md, err := client.Dataset("my_dataset").Update(ctx, bigquery.DatasetMetadataToUpdate{Name: "blind"}, "")
	if err != nil {
		// TODO: Handle error.
	}
	fmt.Println(md)
}
readModifyWrite
package main

import (
	"context"
	"fmt"

	"cloud.google.com/go/bigquery"
)

func main() {
	ctx := context.Background()
	client, err := bigquery.NewClient(ctx, "project-id")
	if err != nil {
		// TODO: Handle error.
	}
	ds := client.Dataset("my_dataset")
	md, err := ds.Metadata(ctx)
	if err != nil {
		// TODO: Handle error.
	}
	md2, err := ds.Update(ctx,
		bigquery.DatasetMetadataToUpdate{Name: "new " + md.Name},
		md.ETag)
	if err != nil {
		// TODO: Handle error.
	}
	fmt.Println(md2)
}

DatasetAccessEntry

type DatasetAccessEntry struct {
	// The dataset to which this entry applies.
	Dataset *Dataset
	// The list of target types within the dataset
	// to which this entry applies.
	//
	// Current supported values:
	//
	// VIEWS - This entry applies to views in the dataset.
	TargetTypes []string
}

DatasetAccessEntry is an access entry that refers to resources within another dataset.

DatasetIterator

type DatasetIterator struct {
	// ListHidden causes hidden datasets to be listed when set to true.
	// Set before the first call to Next.
	ListHidden bool

	// Filter restricts the datasets returned by label. The filter syntax is described in
	// https://cloud.google.com/bigquery/docs/labeling-datasets#filtering_datasets_using_labels
	// Set before the first call to Next.
	Filter string

	// The project ID of the listed datasets.
	// Set before the first call to Next.
	ProjectID string
	// contains filtered or unexported fields
}

DatasetIterator iterates over the datasets in a project.

func (*DatasetIterator) Next 

func (it *DatasetIterator) Next() (*Dataset, error)

Next returns the next Dataset. Its second return value is iterator.Done if there are no more results. Once Next returns Done, all subsequent calls will return Done.

Example

package main

import (
	"context"
	"fmt"

	"cloud.google.com/go/bigquery"
	"google.golang.org/api/iterator"
)

func main() {
	ctx := context.Background()
	client, err := bigquery.NewClient(ctx, "project-id")
	if err != nil {
		// TODO: Handle error.
	}
	it := client.Datasets(ctx)
	for {
		ds, err := it.Next()
		if err == iterator.Done {
			break
		}
		if err != nil {
			// TODO: Handle error.
		}
		fmt.Println(ds)
	}
}

func (*DatasetIterator) PageInfo 

func (it *DatasetIterator) PageInfo() *iterator.