BigQuery console guide: BigQuery Studio, the query editor and job history

How to open the BigQuery console

Google’s steps (read in October 2026):

1. Sign in to the Google Cloud console and select a project on the project selector. Selecting a project needs no specific role, only a role on that project. Creating one needs Project Creator (roles/resourcemanager.projectCreator).
2. Open Navigation menu > Solutions > All products, and in the Analytics section click BigQuery. Or go straight to https://console.cloud.google.com/bigquery.
3. For new projects the BigQuery API is already enabled. Billing is optional: without it you work in the BigQuery sandbox.

If the Explorer loads slowly because a project has many datasets, Google suggests adding ?minimal=true to the URL so the resource tree doesn’t preload.

The navigation menu

The BigQuery navigation menu sits on the far left. Expand or collapse it with Toggle BigQuery navigation menu. As of October 2026 it has four pages and a set of sections:

• Overview: tutorials and features, filterable by role (Data analysis, Data science, Data engineering, Data administration).
• Studio: the console itself.
• Search (Preview): natural-language search for Google Cloud resources.
• Agents (Preview): data agents that answer questions about BigQuery tables.
• Pipelines and integration: data transfers, Dataform repositories and scheduled resources such as scheduled queries.
• Governance: shared data exchanges, data clean rooms, the Security center, metadata.
• Administration: monitoring, jobs, capacity, disaster recovery, recommendations.
• Migration, Partner Center, Settings and Release notes.

Older guides list “Analytics Hub”, “SQL Translation” and “BI Engine” as top-level items. In the current menu those jobs live inside the Governance, Migration and Administration sections.

The left pane: Explorer, Classic Explorer, Files and Repository

The left pane of the Studio page has four tabs:

• Explorer: tables, views, routines and other resources, a search box, Add data, and Job history. Expand a project with Toggle node, then click Datasets to list them in a new tab.
• Classic Explorer: the legacy version of the Explorer pane. Classic saved queries still live here, in the (Classic) Queries folder.
• Files (Preview): user and team folders for code assets such as saved queries and notebooks.
• Repository (Preview): Git version control for files you use in BigQuery.

Click a dataset and you get an Overview tab (its tables, models and routines) and a Details tab (metadata). Click a table and you get its schema, details and a Preview tab. Preview is the cheap way to look at rows: Google’s cost guide says “Don’t run queries to explore or preview table data” and that preview lets you “view data at no charge and without affecting quotas” (read in October 2026). For more on what each table tab shows, see the table explorer guide and BigQuery schema.

Trap: the Explorer shows every project you have a role on. A query runs in the project selected at the top of the console and is billed there, whatever table you clicked.

Tabs, split view and the Home tab

Every resource you open and every new query is a tab in the details pane. Google documents three behaviours worth knowing:

• Tabs get reused. “To reduce tab proliferation, clicking a resource opens it within the same tab.” Ctrl-click (Command-click on macOS) or double-click the tab name to keep it; the name changes from italic to regular. Recent tabs brings back one you replaced by accident.
• Split view. From a tab’s Open menu, choose Split tab to left or Split tab to right. From a table, Query > In split tab opens the editor next to the schema, and clicking a field name inserts it into the query.
• Home tab. Create new opens a SQL query, notebook, Apache Spark notebook, data canvas, data preparation file, pipeline or table. Recent lists your 10 most recently accessed resources.

Keyboard shortcuts are listed under BigQuery Studio shortcuts in the Studio toolbar.

The query editor

Click SQL query (on the Home tab, or the add icon in the details pane) to open an editor tab named Untitled query. Next to it the Reference panel lists recent and starred tables and views; it can insert a query snippet or a field name for you. Tools > Parser-based auto-completion turns autocomplete on or off.

The validator and the bytes estimate

As you type, the editor validates the query. Google’s wording: “If the query is valid, then a check mark automatically appears along with the amount of data that the query will process. If the query is invalid, then an exclamation point appears along with an error message” (Run a query, read in October 2026). That estimate is a dry run, and “you are not charged for performing a dry run.”

The same dry run is available outside the console. These are its results for three queries on a public table:

SELECT * FROM `bigquery-public-data.usa_names.usa_1910_2013`;

SELECT * FROM `bigquery-public-data.usa_names.usa_1910_2013` LIMIT 10;

SELECT name, SUM(number) AS total
FROM `bigquery-public-data.usa_names.usa_1910_2013`
GROUP BY name
ORDER BY total DESC
LIMIT 5;

Trap: LIMIT did not lower the estimate at all. BigQuery reads whole columns, so the cost follows the columns you name, not the rows you return. Google says the same: for non-clustered tables “a LIMIT clause won’t reduce compute costs.” Naming two of the five columns halved the bytes.

The estimate also shows how much a filter on a date-sharded table saves. On the public GA4 sample export, which has 92 daily tables from events_20201101 to events_20210131:

SELECT event_name, COUNT(*) AS events
FROM `bigquery-public-data.ga4_obfuscated_sample_ecommerce.events_*`
WHERE _TABLE_SUFFIX = '20210131'
GROUP BY event_name
ORDER BY events DESC
LIMIT 5;

When the one-day query ran, it processed 339,894 bytes but billed 10,485,760. That is the pricing rule, not a bug: “Charges are rounded up to the nearest MB, with a minimum 10 MB data processed per table referenced by the query” (BigQuery pricing, read in October 2026). On demand, the first 1 TiB a month is free and the rest is $6.25 per TiB. More in BigQuery pricing and the GA4 to BigQuery guide.

Query settings

Edit > Query settings sets, among others: a destination table, job priority, whether to use cached results, a job timeout, session mode, maximum bytes billed, SQL dialect, location and reservation. Maximum bytes billed is the one to set on an unfamiliar table: a query that would bill more fails instead of running.

Trap: location. Google’s page on running queries says the location “must match that of the dataset(s) referenced in the query.” A dataset’s location is fixed when it is created and can’t be changed later.

The query results pane

Click Run. With no destination table, “the query job writes the output to a temporary (cache) table.” The Query results pane opens under the editor. Google’s docs (read in October 2026) describe these tabs:

• Results: the rows. Open sort menu on a column sorts them; if the sort itself would process bytes, the menu says how many.
• Visualization: a chart of the result. Line, Bar and Scatter take numeric, date, time and string dimensions with numeric measures; a GEOGRAPHY column defaults to a map. Charts need at least one numeric column.
• JSON: the result as JSON, column names as keys.
• Job information: details of the job, including performance insights for queries that regressed.
• Execution details and Execution graph: the query plan. The graph shows stages as nodes; stages with an information icon have performance insights, and Highlight top stages by duration or by processing points at the slow part.

For a record of what ran, use Job history (next section).

Saving results

From the results pane you can save to a local file, Google Sheets or Google Drive. Google’s limits (read in October 2026): local downloads are CSV or newline-delimited JSON, at most 10 MB for a CSV; Drive takes CSV or JSON up to 1 GB; results with nested or repeated fields can’t go to Sheets. For anything larger, save to a table. Details in export BigQuery to CSV.

Job history: your own audit trail

“Every time you load, export, query, or copy data, BigQuery automatically creates, schedules, and runs a job.” In the Explorer tab click Job history, then Personal history for your jobs or Project history for everyone’s. Actions > View job in editor reopens the exact SQL. Google: “BigQuery saves a six-month job history for all the jobs of a project, for all locations.”

The same history is queryable, which is how you prove what ran. We ran the five-name query above twice, with a label, and then asked JOBS_BY_USER for those two jobs:

SELECT
  creation_time,
  statement_type,
  cache_hit,
  total_bytes_processed,
  total_bytes_billed
FROM `region-us`.INFORMATION_SCHEMA.JOBS_BY_USER
WHERE creation_time > TIMESTAMP_SUB(CURRENT_TIMESTAMP(), INTERVAL 30 MINUTE)
  AND (SELECT value FROM UNNEST(labels) WHERE key = 'article') = 'bigquery-user-interface'
ORDER BY creation_time;

The second run returned the same five names (James, 4,942,431 first) from the cache and billed nothing. Billed bytes on the first run are higher than processed because of the round-up to the nearest MB.

A weekly check of your own usage:

SELECT
  COUNT(*) AS queries,
  COUNTIF(cache_hit) AS from_cache,
  ROUND(SUM(total_bytes_billed) / POW(1024, 3), 2) AS gib_billed
FROM `region-us`.INFORMATION_SCHEMA.JOBS_BY_USER
WHERE creation_time > TIMESTAMP_SUB(CURRENT_TIMESTAMP(), INTERVAL 7 DAY)
  AND job_type = 'QUERY';

That is our test account’s week, most of it queries run by an AI agent, which is exactly why the view is worth having.

Traps: the region qualifier matters (region-us sees only US jobs), and the view keeps “job history for the past 180 days.” The history query is not free either: the weekly one billed 30 MiB. For more ways to read the log, see BigQuery query history.

Saved queries

Save > Save query (or Ctrl+S, Command+S on macOS) names the query and creates its first version. Saved queries are now “BigQuery Studio code assets powered by Dataform”: they keep a version history you can revert to, are shared through IAM, and appear in the Files tab. Two things from Google’s docs (read in October 2026) surprise people:

• Anyone with the dataform.repositories.list permission, which BigQuery Job User, BigQuery Studio User and BigQuery User include, “can see all code assets in the Explorer panel” of the project, shared or not.
• People you share a query with “only see the most recent version”; autosaved changes you haven’t saved don’t reach them.

Classic saved queries still exist under Save Query (Classic); Google says their “deprecation timeline is being reviewed.”

Scheduled queries

Run the query, then File > Schedule. In New scheduled query give it a name, pick Repeats (daily by default, or Custom with a spec such as every 6 hours), and for a SELECT set a destination dataset and table. Scheduled queries use the BigQuery Data Transfer Service; you find them under Scheduling in the navigation menu, filtered to Scheduled Query. Google’s notes: “The minimum duration between scheduled queries is 5 minutes,” and runs exactly on the hour “might trigger multiple times,” so use 08:58 rather than 09:00.

Notebooks, data canvas and Gemini

• Notebooks. Colab Enterprise notebooks run inside BigQuery Studio and mix SQL, Python, text and charts, with BigQuery DataFrames available without setup.
• Data canvas. A Gemini in BigQuery feature that lets you “find, transform, query, and visualize data by using natural language prompts” on a graph of steps. Google adds that it “isn’t intended for direct use by business users.”
• Gemini in the editor. The Gemini menu in the toolbar switches on auto-completion (Preview), SQL from a comment, the SQL generation tool, and query explanation. Highlight a query and choose Explain this query; highlight an error and choose Refine > Fix it (Preview). It needs the Gemini for Google Cloud User role, and “Only English language prompts are supported.”

Whatever Gemini writes runs as ordinary SQL, so the same validator, bytes estimate and job history apply to it.

Creating a dataset

In the Explorer, select the project, click View actions > Create dataset, enter a Dataset ID, and pick a Location type. Google’s warning: “After a dataset is created, the location can’t be changed.” You need bigquery.datasets.create, which BigQuery Data Editor, Data Owner, User and Admin include. See BigQuery IAM roles.

Where GA4 export data appears

If GA4 is linked to BigQuery, its export lands in a dataset named analytics_<property_id>, with a daily events_YYYYMMDD table and, with streaming export, events_intraday_YYYYMMDD (GA4 BigQuery Export schema, read in October 2026). One table per day is why the query above reads events_* and filters on _TABLE_SUFFIX. How to read the nested event rows is covered in query GA4 event data in BigQuery.

From a console query to a number people reuse

The console is where a query gets written, estimated and checked. It is not where other people should get the number: a saved query shared by link gets copied into a sheet, edited, and the next report disagrees with it.

In OWOX Data Marts the query you validated in the console is published as a Data Mart, with a description of what each column means, and delivered on a schedule to Google Sheets, Data Studio, Excel or an AI assistant. The SQL behind every delivered number stays one click away, and OWOX Data Marts runs it as written in your own BigQuery project; it does not transform the data. See how that works for spreadsheets and for a shared data model.

Google BigQueryBigQuery CREATE TABLE and DROP TABLE: Syntax and ExamplesOctober 16, 2026→Google BigQueryBigQuery ALTER TABLE: Add, Rename, Drop Columns, Change TypesOctober 6, 2026→Google BigQueryBigQuery DML: INSERT, UPDATE, DELETE and MERGE with Tested ExamplesOctober 16, 2026→

See all articles →