BigQuery Queries

What You Can Ask

Just describe what you need in natural language:

• "Show me the top 10 customers by revenue this month"
• "What tables are in the analytics dataset?"
• "Describe the schema of the orders table"
• "Compare this week's signups to last week"
• "Create a chart of daily revenue for the past 30 days"

Available Tools

run_query

Executes a BigQuery SQL query and returns results.

• Default limit: 1,000 rows (adjustable)
• Large results (>50KB) are automatically saved to a session database for follow-up queries
• Includes bytes_processed for cost awareness
• Results formatted as markdown tables for small datasets

list_datasets

Lists all datasets in the configured BigQuery project.

list_tables

Lists all tables and views in a specific dataset, including their type (TABLE, VIEW, SNAPSHOT).

get_table_schema

Returns detailed schema information for a table:

• Column names, types, and descriptions
• Row count and table size
• Creation date

Working with Large Results

When a query returns more data than fits in a chat message, the bot automatically saves the full result to a temporary session database. You can then:

• Ask follow-up questions about the saved data
• Request aggregations or filters on the cached result
• Export it as a CSV or chart

Cost Awareness

Every query reports how many bytes were processed. BigQuery charges per byte scanned, so the bot can help you estimate costs. Tips:

• Always use LIMIT when exploring unfamiliar tables
• Select only the columns you need
• Use partitioned and clustered tables when available
• The bot will suggest optimizations when it detects expensive queries

Charts and Exports

After running a query, you can ask the bot to visualize the results as a chart (line, bar, pie, scatter, etc.) or export them as CSV or PDF.