Skip to main content
Skip table of contents

Google BigQuery

Google BigQuery is a fully managed cloud data warehouse for analytics. While Google BigQuery works in conjunction with Google Storage for interactive analysis of massively large data sets it can scan TeraBytes in seconds and PetaBytes in minutes.

Type Name

bigquery

Connection Properties

Template name: bigquery

Appropriate translator name: bigquery

Properties:

  • projectId (required)
  • transformQuery (default: TRUE); (obsolete)

  • credentialFactory (default: com.datavirtuality.dv.core.oauth.credential.BigQueryOAuthCredentialFactory)

  • allowLargeResults (obsolete)
  • largeResultsDefaultDataset
  • tableMaxResults
  • fetchSize

  • refreshToken

  • accessToken

  • expirationTimeMilliseconds

  • region

  • authCode

  • redirectUri
  • user-name (required)
  • password (default: empty)
  • ClientId
  • ClientSecret
  • driver (default: bigquery)
  • driver-class (default: com.datavirtuality.jdbc.bq.BQDriver)
  • storageProjectId (default: empty)
  • storageUser (default: empty)
  • storagePassword (default: empty)
  • new-connection-sql
  • check-valid-connection-sql (default: select 1)
  • min-pool-size (default: 2)
  • max-pool-size (default: 70)
  • readTimeout (default: 20000, 0 for an infinite, a negative for the default)
  • connectTimeout (default: 20000, 0 for an infinite, a negative for the default)

Here is an example:

SQL 
CALL SYSADMIN.createConnection('bq2','bigquery','projectId=XXXX,user-name=XXXX@developer.gserviceaccount.com,password=$${jboss.server.config.dir}/../deployments/bigquery.p12') ;;
CALL  SYSADMIN.createDatasource('bq2','bigquery','importer.useFullSchemaName=false,importer.useCatalogName=false','supportsNativeQueries=true,uploadMode=CLOUDSTORAGE,bucketName=dv_upload_bucket,useDdl=false') ;;

It is highly recommended to specify a bucketName when creating a BigQuery data source because of the following:

  • a bucket creation limit that can be exceeded while performing multiple data inserts. Please refer to the Google Cloudstorage documentation for detailed information;
  • region mismatch between a BigQuery data set and a temporary bucket which may cause errors while inserting data. Please, refer to the Google BigQuery documentation for more details.

Translator Properties

Translator Properties Shared by All JDBC Connectors

(Properties listed in alphabetical order)

To view the full table, click the expand button in its top right corner


Name

Description

Default value

comparisonStringConversion

Sets a template to convert WHERE clauses on strings

Examples

SQL 
comparisonStringConversion=%s
-- no conversion will be applied

comparisonStringConversion=binary %s
-- WHERE binary someStringExpression LIKE someOtherStringExpression

comparisonStringConversion=(%s COLLATE Latin1_General_CS_AS)
-- WHERE (someStringExpression COLLATE Latin1_General_CS_AS LIKE someOtherStringExpression)


  • MS SQL Server:
    (%s COLLATE Latin1_General_CS_AS)
  • MySQL, SingleStore:
    binary %s
  • Others:
    %s

DatabaseTimeZone

Database time zone, used when fetching date, time, or timestamp values

System default time zone

DatabaseVersion

Specific database version, used to fine-tune pushdown support

Automatically detected by the server through the data source JDBC driver, if possible

forbidNonMatchingNamesOnly considered when importer.tableNamePattern is set. When set to FALSE, allows creation of tables with names that do not match the tableNamePattern. Tables with non-matching names will only be accessible until the server restartsTRUE
ForceQuotedIdentifiers

If TRUE, all identifiers in the final query (that is being generated by a translator to be sent directly to the target DBMS) will be quoted

TRUE

MaxPreparedInsertBatchSize

Maximum size of prepared insert batch

2048

OrderByStringConversion

Sets a template to convert ORDER BY clauses on strings

Examples

SQL 
OrderByStringConversion=%s
-- no conversion will be applied

OrderByStringConversion=(binary %s)
-- ORDER BY (binary someStringExpression)

OrderByStringConversion=cast((%s) as varchar(2000) ccsid 1208)
-- ORDER BY cast((someStringExpression) as varchar(2000) ccsid 1208)
  • DB2 for AS/400:
    cast((%s) as varchar(2000) ccsid 1208)
  • DB2 for zOS:
    cast((%s) as varchar(2000) ccsid ASCII)
  • MS SQL Server:
    (cast(cast(%s as varchar) as varbinary)
  • MySQL, SingleStore:
    (binary %s)
  • PostgreSQL:
    (%s COLLATE \"C\")
  • Others:
    %s
supportsConvertFromClobToString If TRUE, indicates that the translator supports the CONVERT/CAST function from clob to string
  • MySQL, PostgreSQL, Redshift, Greenplum, Oracle, MS SQL Server:
    TRUE 
  • Others:
    FALSE

supportsNativeQueries

Forces a translator to issue a native() system procedure that can be used to pass native queries directly to an underlying DBMS

Example

SQL 
SELECT x.* 
FROM table
  ( 
    CALL "dwh.native"("request" => 'SELECT query, pid, elapsed, substring FROM svl_qlog ORDER BY STARTTIME DESC LIMIT 200') 
  ) w
  , ARRAYTABLE( w.tuple COLUMNS query string, pid integer , elapsed string, "substring" string ) x;

More usage examples

  • MySQL, PostgreSQL, Redshift, Snowflake, Oracle, MS SQL Server; Exasol, Vertica, Bigquery, SingleStore, Azure:
    TRUE
  • Others:
    FALSE
supportsOrderByAlias

If FALSE, the CData Virtuality Server does not generate aliases in the ORDER BY clause.

If TRUE, indicates that the source supports aliases in the ORDER BY clause

  • PostgreSQL:
    TRUE
  • Others:
    • if OrderByStringConversion is not set, same as supportsOrderBy
    • if OrderByStringConversion is set, default for supportsOrderByAlias is  FALSE
supportsOrderByString

If FALSE, the CData Virtuality Server does not push down the ORDER BY clause if it contains at least one expression of string type.

If TRUE, indicates that the source supports strings in the ORDER BY clause

TrimStrings

If TRUE, trims trailing whitespace from fixed-length character strings. Please note that the CData Virtuality Server only has string or varchar type that treats trailing whitespace as meaningful

FALSE

UseBindVariables

if TRUE, indicates that PreparedStatements should be used and that literal values in the source query should be replaced with bind variables. If FALSE, only LOB values will trigger the use of PreparedStatements

TRUE

UseCommentsInSourceQuery

Embeds a / comment / leading comment with session/request id in the source SQL query for informational purposes

FALSE

The names of the translator properties are case-sensitive.

Translator Properties Specific for Google BigQuery

To view the full table, click the expand button in its top right corner

Name

Description

Default value

replaceNullCharsWith

String property. If set, the translator replaces all null characters in strings before executing INSERT / UPDATE with the string specified as the property value. You may specify an empty string as wellSingle space

uploadMode

Values:  CLOUDSTORAGE (default), STREAMING, INSERTS

  • CLOUDSTORAGE mode uses a Google Cloud Storage bucket to upload data. The bucket can be configured via bucketName or bucketPrefix translator parameters. If no bucketName is specified, the bucket will be created;
  • STREAMING mode streams data via the Big Query API;
  • INSERTS mode uses JDBC prepared statements for inserting data
CLOUDSTORAGE

bucketName

Value: bucket name

Only for the default (CLOUDSTORAGE) upload mode.
Specifies a bucket name to upload data files to. If specified, the specified bucket is assumed to exist (no creation/deletion operations performed). The region specified for the bucket must be the same as the region specified via the region translator property. If no bucketName is specified, a bucket will be created and deleted after the operation is finished.


bucketPrefix

Value: bucket prefix

Only for the default (CLOUDSTORAGE) upload mode.
Specifies a prefix of a bucket name to upload data files to. The name of the bucket created for uploading will start with the specified prefix. The bucket will be deleted after the operation is finished.


useDdl

Value: boolean

If TRUE, creates and drops are performed using JDBC statements instead of API

FALSE

setACLforBucket

If set to TRUE, ACL roles are set for the elements of the bucket used by the data source.

If set to FALSE, only IAM permissions are used. Thus, it should be set when uniform bucket-level access is configured for the bucket

TRUE

Data Source Properties

Data Source Properties Shared by All JDBC Connectors

(Properties listed in alphabetical order)

To view the full table, click the expand button in its top right corner


Name

Description

Default

importer.autoCorrectColumnNames Replaces . in a column name with _ as the period character is not supported by the CData Virtuality Server in column names TRUE
importer.catalog

Database catalogs to use. Can be used if the db connection property is omitted.

Only for Microsoft SQL Server and Snowflake:

  • Multiple catalogs (multiple databases) usage is supported;
  • '%' as a wildcard is allowed: for example, importer.catalog="foo%" will import foo, foobar, etc.
  • To specify several catalog names or/and patterns, values should be comma-separated and enclosed within double quotes: importer.catalog="schema1,schema2,pattern1%,pattern2%";
  • To use all catalogs, the value should be '%': importer.catalog="%"
  • To escape a wildcard in a catalog name, use the '\' escape character. The value should be as follows: importer.catalog="test\_tables"

Exasol:

SQL 
EXA_DB

All others: empty

importer.defaultSchema
  • Only for data sources that are also supported as Analytical Storage (SQLServer, MySQL, Oracle, PostgreSQL, Redshift);
  • When the property is correctly set, SELECT INTO, CREATE, and DROP TABLE commands are enabled for that data source;
  • Must point to the original schema name in the DBMS (e.g. importer.defaultSchema=public);
  • Needs specifying where tables will be created or dropped in the source DBMS;
  • Not meant to be used with Analytical Storage data source

Please note that writing into a data source is only possible if this parameter is set.

Empty
importer.enableMetadataCache Turns on metadata cache for a single data source even when the global option is turned off. Together with importer.skipMetadataLoadOnStartup=true, it allows using materialized views after server restart when the original source is unavailable
 FALSE
importer.excludeProcedures Case-insensitive regular expression that will exclude a matching fully qualified procedure name from import
Empty
importer.excludeSchemas Comma-separated list of schemas (no % or ? wildcards allowed) to exclude listed schemas from import. A schema specified in defaultSchema or schemaPattern will be imported despite being listed in excludeSchemas. Helps to speed up metadata loading

Oracle:

SQL 
APEX_PUBLIC_USER,
DIP,
FLOWS_040100,
FLOWS_020100,
FLOWS_FILES,
MDDATA,
ORACLE_OCM,
SPATIAL_CSW_ADMIN_USR,
SPATIAL_WFS_ADMIN_USR,
XS$NULL,
BI,
HR,
OE,
PM,
IX,
SH,
SYS,
SYSTEM,
MDSYS,
CTXSYS

All others: empty

importer.excludeTables

Case-insensitive regular expression that will exclude a matching fully qualified table name from import. Does not speed up metadata loading. Here are some examples:

1. Excluding all tables in the (source) schemas sys and INFORMATION_SCHEMA:

SQL 
importer.excludeTables=(.*[.]sys[.].*|.*[.]INFORMATION_SCHEMA[.].*)

2. Excluding all tables except the ones starting with "public.br" and "public.mk" using a negative lookahead: 

SQL 
importer.excludeTables=(?!public\.(br|mk)).*

3. Excluding "tablename11" from the list ["tablename1", "tablename11", "company", "companies"]:

SQL 
importer.excludeTables=.*\.(?!\btablename1\b|\bcompan).*
  • ".*\." will match the schema portion of the table name;
  • "\b" is a word boundary. It is used to define the boundary between text characters and non-character text;
  • "\btablename1\b" will only match tablename1 and not tablename11 because of the use of "\b";
  • "\bcompan" will match "company" and "companies" because we did not use "\b" at the end to delimit a word boundary;
  • ".*" at the end is necessary to match the trailing characters after the name

Empty

importer.fetchSize

Fetch size assigned to a resultset on loading metadata

No default value

importer.importApproximateIndexes

If set to TRUE, imports approximate index information

TRUE

importer.importIndexes If set to TRUE, imports index/unique key/cardinality information FALSE
importer.importKeys If set to TRUE, imports primary and foreign keys FALSE

importer.importProcedures

If set to TRUE, imports procedures and procedure columns. Overloaded procedures can be imported with an additional option useProcedureSpecificName.

Please note that it is currently not possible to import procedures which use the same name for more than one parameter (e.g. same name for IN and OUT parameters). Such procedures can be excluded from import with the parameter excludeProcedures.
Please note that it is not always possible to import procedure result set columns due to database limitations

FALSE

importer.loadColumnsTableByTableSet to TRUE to force table by table metadata processingFALSE/TRUE only for Netsuite and SAP Advantage Database Server
importer.loadMetadataWithJdbc If set to TRUE, turns off all custom metadata load ways FALSE
importer.procedureNamePattern Procedure(s) to import. If omitted, all procedures will be imported. % as a wildcard is allowed: for example, importer.procedurePatternName=foo% will import foo , foobar, etc. W orks only in combination with importProcedures Empty

importer.quoteNameInSource

If set to FALSE, directs the Data Virtuality Server to create source queries using unquoted identifiers

TRUE

importer.renameDuplicateColumns If set to TRUE,  renames duplicate columns caused by either mixed case collisions or autoCorrectColumnNames replacing . with _. The suffix _n where n is an integer will be added to make the name unique TRUE
importer.renameDuplicateTables If set to TRUE, renames duplicate tables caused by mixed case collisions. The suffix _n where n is an integer will be added to make the name unique TRUE
importer.replaceSpecSymbsInColNamesIf set to TRUE, replaces all special symbols (any symbols not in the ^A-Za-z0-9_ sequence) to the _ symbol in column names of tables FALSE / TRUE only for BigQuery
importer.schemaPattern Schema(s) to import. If omitted or has "" value, all schemas will be imported. % as wildcard is allowed: for example, importer.schemaPattern=foo% will import foo, foobar, etc. To specify several schema names or/and patterns, values should be comma-separated and enclosed within double quotes: importer.schemaPattern="schema1,schema2,pattern1%,pattern2%". For proper escaping of special characters depending on the type of data source, check Escaping special characters in schema names or use wildcards instead: "[schema_name]" can be rewritten as "%schema%name%". Helps to speed up metadata loadingEmpty
importer.skipMetadataLoadOnStartup

If set to TRUE, allows skipping metadata loading on server startup if there's cached metadata for the data source. . Together with importer.enableMetadataCache=true, it allows using materialized views after server restart when the original source is unavailable

 FALSE
importer.tableNamePattern Table(s) to import. If omitted, all tables will be imported. % as a wildcard is allowed: for example, importer.tableNamePattern=foo% will import foo, foobar, etcEmpty
importer.tableTypes

Comma-separated list (without spaces) of table types to import. Available types depend on the DBMS. Usual format: "TABLE,VIEW".

Other typical types are "SYSTEM TABLE", "GLOBAL TEMPORARY", "LOCAL TEMPORARY", "ALIAS", "SYNONYM"

Empty
importer.useCatalogName If set to TRUE, uses any non-null/non-empty catalogue name as part of the name in source, e.g. "catalogue"."table"."column", and in the CData Virtuality Server runtime name if useFullSchemaName is TRUE. If set to FALSE, will not use the catalogue name in either the name in source or the CData Virtuality Server runtime name. Should be set to FALSE for sources that do not fully support a catalogue concept, but return a non-null catalogue name in their metadata - such as HSQL TRUE / FALSE only for Hive and EXASOL
importer.useFullSchemaName

If set to FALSE, directs the importer to drop the source catalogue/schema from the CData Virtuality Server object name so that the CData Virtuality Server fully qualified name will be in the form of <model name>.<table name>.

Please note that this may lead to objects with duplicate names when importing from multiple schemas, which results in an exception

 TRUE
importer.useProcedureSpecificName If set to TRUE, allows the import of overloaded procedures (which will normally result in a duplicate procedure error) by using the unique procedure specific name in the CData Virtuality Server. This option will only work with JDBC 4.0 compatible drivers that report specific names FALSE
importer.widenUnsignedTypes If set to TRUE, converts unsigned types to the next widest type. For example, SQL Server reports tinyint as an unsigned type. With this option enabled, tinyint would be imported as a short instead of a byte TRUE
The names of the data source properties are case-sensitive.

Creating/Dropping Tables and Inserting Data

In order to create/drop tables and insert data into a BigQeury data source, the importer.defaultSchema  data source property should be set to the name of the target dataset. Here is an example:

SQL 
CALL SYSADMIN.createConnection(name => 'bq', jbossCliTemplateName => 'bigquery', connectionOrResourceAdapterProperties => 'projectId=XXXX,user-name=XXXX-ZZZZ@developer.gserviceaccount.com,password=PATH_TO_KEY_FILE') ;;
CALL SYSADMIN.createDatasource(name => 'bq', translator => 'bigquery', modelProperties => 'importer.useCatalogName=false,importer.defaultSchema=TARGET_DATA_SET,importer.schemaPattern=SCHEMA_PATTERN_INCLUDING_TARGET_DATASET,importer.useFullSchemaName=false') ;;

Segmentation and Optimization

The BigQuery partitioning functionality is supported in both API and DDL modes.

API Mode

In the API mode (when the useDdl translator property is set to FALSE), tables can be partitioned on the DATE or TIMESTAMP column. On the BigQuery side, such tables will be partitioned with daily granularity:

SQL 
CREATE TABLE bigquery.partition1 (i INTEGER, d DATE, s STRING) OPTIONS (partition_by 'd') ;;

DDL Mode

In DDL mode (when the useDdl translator property is set to TRUE), partitioning can be configured via the BigQuery syntax, including integer-range partitioning:

1.. Creating a table partitioned on a date range:

SQL 
CREATE TABLE bigquery_ddl.partition_ddl_date (i integer, d timestamp, s string) OPTIONS (partition_by 'DATE(d)') ;;

2. Creating a table partitioned on integer range:

SQL 
CREATE TABLE bigquery_ddl.partition_ddl_integer (i integer, d timestamp, s string) OPTIONS (partition_by 'RANGE_BUCKET(i, GENERATE_ARRAY(0, 100, 10))') ;;

Clustering

Clustering is supported via the OPTIONS clause:

SQL 
CREATE TABLE bigquery.cluster (i integer, d date, ts timestamp, s string, b boolean) OPTIONS (cluster_by 'i,d,ts,s') ;;

Configuring 3-legged OAuth for BigQuery

To set up 3-legged OAuth with BigQuery, follow these steps:

1. Log in to the Developer Console for your BigQuery project.

2. Generate new OAuth2 credentials (a pair of CLIENT_ID and CLIENT_SECRET keys).

3. Substitute CLIENT_ID in the link below with your CLIENT_ID and follow this link in the browser:

https: //accounts.google.com/o/oauth2/auth?scope=https%3A%2F%2Fwww.googleapis.com%2Fauth%2Fbigquery&access_type=offline&state=security_token&response_type=code&client_id=CLIENT_ID&approval_prompt=force&redirect_uri=http%3A%2F%2Flocalhost%3A9000%2Fredirect.html%3Fto%3Dconnect/bigquery/oauth

4. Choose 'Allow' to provide access:

5. You will be redirected to a non-existent page, which is correct. Copy the page URL to obtain the code:

(warning) The OAuth code is only valid for 10 minutes

http: //localhost:9000/redirect.html?to=connect/bigquery/oauth&state=security_token&code=4/cwHserUpJwJSbKWQX-1Ua-cw5xWtdJPEDBx4nfbRmJ9F0wW2NVlLEhE8umS9bE6cdXas0IK9FZfirA02ZYDeTTw&scope=https://www.googleapis.com/auth/bigquery


6. Create a data source using the CLIENT_ID and CLIENT_SECRET obtained in step 3, AUTH_CODE obtained in step 5, and PROJECT_ID of your BigQuery project. The redirectUrl should be the same as in steps 3 and 5:

SQL 
CALL SYSADMIN.createConnection('bq3legged','bigquery','authType=OAUTH,
projectId=PROJECT_ID ,
authCode=AUTH_CODE ,
redirectUri=http%3A%2F%2Flocalhost%3A9000%2Fredirect.html%3Fto%3Dconnect/bigquery/oauth,
user-name=CLIENT_ID,
password=CLIENT_SECRET') ;;
CALL SYSADMIN.createDatasource('bq3legged','bigquery','importer.schemaPattern=bq_schema_pattern,importer.defaultSchema=bq_schema,
	importer.useFullSchemaName=false,importer.useCatalogName=false','supportsNativeQueries=true') ;;

See Also

Partitioning Tables in BigQuery to learn how to convert a non-partitioned table into a partitioned one for improved query performance


JavaScript errors detected

Please note, these errors can depend on your browser setup.

If this problem persists, please contact our support.

Contact Support Close