Release Notes

This page documents production updates to BigQuery. We recommend that BigQuery developers periodically check this list for any new announcements.

Subscribe to the BigQuery release notes. Subscribe

May 15, 2018

Support for the NUMERIC data type is now in Beta.

May 4, 2018

The following quotas have changed:

May 1, 2018


Updated version of the Simba JDBC driver.

BigQuery Data Transfer Service

When you create an AdWords transfer, the maximum value for the refresh window is now 7 days. Existing transfers are unaffected by this change.

April 17, 2018

BigQuery is now available in the Tokyo (asia-northeast1) region.

BigQuery now includes permissions. still allows users to list all jobs in a project, but details and metadata are redacted for jobs submitted by other users.

April 13, 2018

By default, the number of export jobs per project per day has increased from 1,000 to 50,000.

April 10, 2018

Support for DML MERGE statements is now in Beta.

April 6, 2018

BigQuery Data Transfer Service

DoubleClick for Publishers transfers now support API version v201802.

March 30, 2018

BigQuery standard SQL now supports: the ERROR function to return errors and the SAFE prefix to suppress errors.

BigQuery now allows you to monitor slot usage by job type in Stackdriver.

March 5, 2018

BigQuery data definition language now supports creating a table from the result of a query.

BigQuery support for using DML statements with partitioned tables is now in Beta.

March 1, 2018

Support for loading Parquet files is now in Beta.

Updated version of Simba JDBC driver.

February 16, 2018

BigQuery Data Transfer Service

Google AdWords transfers now support manager accounts (MCCs).

AdWords transfers now support API version 201710.

February 15, 2018

BigQuery Data Transfer Service

BigQuery Data Transfer Service run notifications are now in Alpha.

February 8, 2018

Several API methods will begin including a location field in their responses: jobs.insert, jobs.get, jobs.cancel, jobs.query, jobs.getQueryResults, datasets.list, and jobs.list. The location field is currently not required in requests.

For example, the id property will include a location in the following format: "id": "[PROJECT_ID]:[LOCATION].[JOB_ID]". The “id” property should continue to be treated as opaque. Also, the "jobReference" property will begin including a location field in the following format:

  "jobReference": {
      "jobId": "[JOB_ID]",
      "location": "[LOCATION]",

Added documentation on manually changing table schemas.

January 17, 2018

BigQuery data definition language support is now in Beta.

January 8, 2018

BigQuery Data Transfer Service

January 5, 2018

ISOWEEK and ISOYEAR are now supported by the following functions: DATE_TRUNC, DATETIME_TRUNC, TIMESTAMP_TRUNC, DATE_DIFF, and DATETIME_DIFF.

December 11, 2017

BigQuery Data Transfer Service

You can now exclude removed or deleted items when setting up a Google Adwords transfer.

You can now check the Configure Jobs option to allow the BigQuery Transfer Service to manage YouTube Channel and Content Owner reporting jobs for you.

Backfill requests for Google Adwords do not overwrite Match Tables.

December 8, 2017

Updated versions of Simba ODBC and JDBC releases.

December 4, 2017

The BigQuery web UI supports additional dimensions for filtering your query history.

November 30, 2017

INTERSECT DISTINCT and EXCEPT DISTINCT are now supported set operators in standard SQL.

FARM_FINGERPRINT is now a supported function in legacy SQL.

WEEK(<WEEKDAY>) is now a valid date part for the EXTRACT function.

There is now a quota of 300 concurrent API requests per user.

Cloud IAM custom roles are now available for BigQuery.

The _PARTITIONTIME pseudo column is now supported when querying data using a wildcard table over partitioned tables.

November 16, 2017

BigQuery Data Transfer Service is now generally available.

November 14, 2017

High compute queries are being phased out so you are no longer required to opt-in. All high compute queries under tier 100 are billed as tier 1. Running a high compute query does not require setting a threshold flag.

All queries above tier 100 will fail with a RESOURCES_EXCEEDED_PER_BYTE error unless the query is running in a reserved instance. Existing users who have run a high compute query higher than tier 100 in the last 30 days retain the ability to run queries at the higher billing tier until October 31, 2018.

November 2, 2017

The IGNORE NULLS and RESPECT NULLS modifiers are now supported in the FIRST_VALUE, LAST_VALUE, and NTH_VALUE analytic functions.

The FOR SYSTEM TIME AS OF expression is now available in the FROM clause.

October 16, 2017

Updated versions of Simba ODBC and JDBC releases.

September 29, 2017

GROUP BY ROLLUP is now supported in standard SQL.

September 28, 2017

The following string functions are now supported in standard SQL:

Added support for creating ingestion-time partitioned tables by specifying partitioning options when you write query results or load data into a new destination table.

September 1, 2017

The following analytic functions are now supported in standard SQL:

August 18, 2017

Added best practices documentation for BigQuery:

August 17, 2017

Updated versions of Simba ODBC and JDBC releases.

August 11, 2017

The following standard SQL statistical aggregate functions now support the OVER clause — COVAR_SAMP, COVAR_POP, and CORR:

August 2, 2017

Announced support for WEEK, ISOWEEK, and ISOYEAR with the standard SQL EXTRACT function:

July 27, 2017

Announced availability of snappy codec for Avro compressed data blocks:

July 26, 2017

Announced availability of standard SQL TO_HEX and FROM_HEX functions:

July 13, 2017

Announced support for using DML statements to modify tables with required fields:

  • DML
  • You can also update a table with required fields by using a SELECT statement and writing to a destination table.

June 19, 2017

June 7, 2017

Announced availability of new public dataset:

June 6, 2017

Announced availability of new public dataset:

June 5, 2017

The following standard SQL statistical aggregate functions now support the OVER clause — STDDEV, STDDEV_POP, STDDEV_SAMP, VAR_POP, VAR_SAMP, VARIANCE:

June 2, 2017

Standard SQL now supports the TO_JSON_STRING function.

May 23, 2017

Added troubleshooting documentation for the BigQuery Data Transfer Service:

May 22, 2017

Announced availability of new public dataset:

May 17, 2017

May 12, 2017

Announced availability of new public datasets:

May 10, 2017

Announced availability of new IAM role:

May 5, 2017

Announced availability of new public dataset:

May 2, 2017

Announced availability of new public datasets:

April 20, 2017

Announced availability of new public dataset:

April 05, 2017

Updated versions of JDBC releases.

March 31, 2017

Announced availability of new public dataset:

March 16, 2017

Updated limits for maximum rows per second when streaming data into BigQuery.

Standard SQL now supports REVERSE and ARRAY_REVERSE functions for reversing STRING, BYTES, and ARRAY types.

March 15, 2017

Standard SQL now supports several functions that use HyperLogLog++ to estimate the number of unique values in a large dataset.

Standard SQL now supports window, or analytic, mode for ARRAY_AGG and STRING_AGG using the OVER clause.

Updated limits for Avro headers and maximum file size for load jobs.

March 14, 2017

Updated versions of JDBC and ODBC releases.

March 9, 2017

Announced beta availability of BigQuery Data Transfer Service.

March 8, 2017

Announced general availability of Data Manipulation Language for nonpartitioned tables in standard SQL.

Announced general availability of IGNORE NULLS and RESPECT NULLS modifiers for ARRAY_AGG and APPROX_QUANTILES in standard SQL.

Announced daily limits for Copying a table in BigQuery.

February 1, 2017

Standard SQL now supports:

Announced support for copying non-partitioned tables to partitioned tables. For more information, see Copying to partitioned tables.

Announced availability of a new public dataset: New York City Tree Census data.

Announced availability of a new public dataset: US Bureau of Labor Statistics data.

December 22, 2016

Announced availability of custom null markers for CSV load jobs.

Standard SQL now supports the following code point functions: TO_CODE_POINTS, CODE_POINTS_TO_BYTES, CODE_POINTS_TO_STRING.

Standard SQL now supports ORDER BY and LIMIT clauses within ARRAY_AGG, ARRAY_CONCAT_AGG, and STRING_AGG.

Hash functions in standard SQL.

URL functions in standard SQL.

Standard SQL now supports parameterized queries, which help prevent SQL injection when queries are constructed using user input.

Announced availability of a new public dataset: New York City Citi Bike Trips data.

Announced availability of a new public dataset: NYC 311 Service Requests data.

Announced availability of a new public dataset: NYPD Motor Vehicle Collisions data.

If you use OMIT...IF to exclude a portion of a record in a repeating field, and the query also selects other independently repeating fields, BigQuery now returns an error. To avoid this error, migrate your query to standard SQL.

November 30, 2016

Announced availability of JSON functions in standard SQL.

Announced availability of a new public dataset: IRS Form 990 Data.

Announced publication of two new tutorials:

November 7, 2016

Announced availability of a new public dataset: Medicare Data.

November 2, 2016

Announced availability of a new public dataset: Open Images Data.

November 1, 2016

Announced JDBC drivers.

You can set a SQL variant, either legacy SQL or standard SQL, by setting a query prefix.

Announced availability of a new public dataset: Major League Baseball Data.

Announced availability of standard SQL safe casting functions, which return a null value in cases where normal casting generates an error.

The BigQuery Client Libraries page now links exclusively to one set of libraries: the Google Cloud Client Libraries. Work is underway to update all sample code throughout the BigQuery documentation to use the Google Cloud Client Libraries.

Bracket notation is now supported for the legacy SQL JSON functions.

Additional documentation is available for access control using Identity and Access Management (IAM):

September 29, 2016

Announced general availability of standard SQL, which is based on the SQL 2011 standard, with extensions for querying nested and repeated data.

New documentation includes:

Announced Data Manipulation Language (DML) functionality for standard SQL, which enables you to insert, update, and delete rows and columns in your BigQuery datasets.

For syntax details and examples, see DML Syntax.

Announced BigQuery Stackdriver Monitoring functionality, which enables you to monitor your BigQuery resources, including creating charts and alerts based on BigQuery metrics.

Announced ODBC drivers.

Flat-Rate Pricing gives you unlimited querying for a fixed monthly price.

You can now save and share queries.

August 26, 2016

The updated beta IAM predefined roles for BigQuery are now accessible to all.

August 1, 2016

When loading Avro files, BigQuery no longer creates an unnamed root-level column named "root." To ease the transition to this new schema, projects with the following contents are exempted from this change through a project override:

  • Projects that contained federated tables based on Avro files as of August 1, 2016
  • Projects that have queried or loaded Avro files in the 30 days prior to August 1, 2016

All affected project owners will be notified and provided the option to remove the project override.

July 27, 2016

Announced support for labeling datasets using the BigQuery API or command-line tool.

June 2, 2016

Announced support for time-partitioned tables.

Announced Beta availability of standard SQL, which is based on the SQL 2011 standard, with extensions for querying nested and repeated data.

Announced a Beta preview of Product Identity and Access Management (IAM) Roles.

May 6, 2016

Query reference updates:

March 23, 2016

December 15, 2015

  • Added Custom quotas, which allow you to set project-wide billing alerts that help prevent runaway costs.
  • Added Query explain, which shows, stage by stage, how BigQuery executes your queries.
  • Added improvements to the Streaming API:
    • You can use the Streaming API in EU locations.
    • You can use Template tables to manage related tables used for streaming.
    • No more warm-up delay. You can now query your data within a few seconds of the first streaming insertion.
  • Added support for analyzing Google Cloud Audit logs that enable you to track usage and access.

August 25, 2015

April 16, 2015

March 25, 2014

  • Added views. A view is a virtual table defined by a SQL query.
  • Added JSON functions.
  • Added table wildcard functions, for a cost-effective way to query data from a specific set of tables.
  • Improved JOIN support by adding CROSS JOIN and multiple JOIN functionality. For more information, see joins.
  • Increased quota limits for streaming data into BigQuery.
  • Added a description property for datasets, tables and fields.
  • Updated the concurrent rate limit and export request quota in the quota policy.
  • Added parallel export capability.
  • Changed the pricing model, lowering rates and adding reserved capacity pricing.
  • Updated batch query functionality. Batch queries no longer have a 30 minute execution delay. Instead, batch jobs now begin as soon as idle resources are available. If the job hasn't started within 3 hours, the job priority changes to interactive.

Sept 18, 2013

June 11, 2013

March 14, 2013

  • Added TIMESTAMP data type functionality.
  • Added GROUP EACH BY functionality for aggregations of more distributed values.
  • Added JOIN EACH functionality for joining tables of any size.
  • Added ability to add columns to table schema after the data is already loaded into BigQuery. To add columns to a table schema, pass the new schema to the schema parameter of Tables.update().

October 1st, 2012


August 28th, 2012


  • New batch queries feature allows for queries to be executed between 30 minutes to 3 hours later, but are charged at a lower cost than interactive queries which execute immediately

New tool release: BigQuery Connector for Excel now allows you to make queries to Google BigQuery from Microsft Excel!

August 17th, 2012


  • CSV chunking - when importing, you can now specify the allowQuotedNewlines flag to be FALSE in the bq command-line tool, or in the API, and the import will be chunked, or processed in parallel. If you have quoted new lines in your data, be sure to set this flag to be TRUE.
  • Deletion of large tables has been fixed
  • Improved validation of project IDs

July 16th, 2012

BigQuery Browser Tool

  • Added syntax highlighting
  • Add a drag handle to the separator below the query editor
  • Fixed some keyboard shortcuts in the query editor
  • New favicon


  • Project owners can now call for jobs created by other users

May 30th, 2012


  • Introduced new quota enforcement schema. User can now have one (1) large query and one (1) pipeline job (import, export, streaming, etc.) concurrently.
  • Added support for leading underscores in field name
  • Improved error handling
  • Added token based pagination support for list operations
    • Clients can use token based pagination to page through a single table snapshot. Subsequent changes to the table are not reflected when paging through list results.
  • Improved list operations for tables and datasets

May 11, 2012


  • Table expiration time is now exposed
  • Better errors for field/rows being too long
  • Custom delimiters added

April 20, 2012

BigQuery Browser Tool

  • New welcome page and more helpful messages when there aren't any queries or jobs to display


  • Clearer quota error messages
  • Copy jobs are now atomic

April 13, 2012

BigQuery Browser Tool

  • Add retries for connection errors
  • Table view: Field names in schemas are left-aligned
  • Table view: Removed "Actions" menu (use the nav panel instead)
  • Table view: Added "Query Table" menu, which prefills query builder with a table query
  • Query builder has simpler prefill logic and a better help bubble


  • Remove

April 5, 2012

BigQuery Browser Tool

  • 10MB limit to uploads done through the BigQuery browser tool
  • Bug fix for zero row results


  • API Discovery docs have typed configuration
  • Bug fix limiting the amount of data returned in a TableData.List
  • Atomic updates of tables for all import jobs
  • Join behavior tightened (no outer join, only left most table can be large)

March 9, 2012

BigQuery Browser Tool

  • Pressing the tab key in the query box now adds a tab instead of tabbing to the next UI element
  • Pressing the enter key in the query box now preserves leading whitespace to the next lines
  • Pressing the enter key no longer starts a query but will always add a newline to the box, and CTRL+ENTER runs the query
  • Navigation panel no longer forgets scroll location and submenu state when jobs are running
  • BigQuery browser tool now has a new URL at; it is still possible to access the browser tool using the old URL

March 1, 2012

BigQuery Browser Tool

  • Faster BigQuery browser tool load time
  • "Max invalid records" support in the BigQuery browser tool


  • Enable support for VARIANCE, STDDEV, and AVG aggregate functions
  • Add job.kind to Jobs:list API response
  • Better error messages for quota-related errors

February 2012

BigQuery Browser Tool

  • 'Show previous query results' option in the query panel
  • The current project view is now highlighted
  • "Refresh" option in the projects drop-down menu


  • Support for loading gzip compressed files from Google Cloud Storage
  • Error handling
    • Faster detection of unreadable files
    • Improvements in handline multibyte UTF-8 characters

November 2011

BigQuery Browser Tool

  • The Web UI now allows JSON schema format to be used in the "Create Table" dialog

    This is useful in situations where developers may want to use the same schema format for API calls, the bq command-line client, and the Web UI.


  • The default configuration.load.createDisposition for load jobs is now CREATE_IF_NEEDED
  • "load" jobs now allow users to specify a number of rows to skip
    • This is useful for skipping header rows or other metadata that is not parseable when you import CSV files.
    • This value applies to all files in the source list.
  • BigQuery can now load data from files that contain delimiters other than the comma
    • Delimiters may be any character expressable as a single Latin-1 byte other than ", \n, and \r.
    • Example (using the bq command line tool): bq load --field_delimiter=: newtable gs://bucket/input.txt

October 2011

BigQuery V2 Release

  • A new REST API
  • A new web user interface
  • Support for JOIN statements
  • Ability to export a table or query result to a CSV file in Google Cloud Storage
  • Support for ACLs on groups of tables (datasets)
  • A new object architecture describing tables, groups of tables, and queries

Send feedback about...