Run one bounded statistics statement over a whole virtual table: its row count and, per named field, distinct count, null fraction and smallest and largest value. It is an ordinary source read: your read access, the source's disclosure rules, metering and outcome reporting apply, and an uncertain outcome is never retried. Later plans use the exact result as analyzed statistics while the source reports the same version; for a source without versions, Explain reports their age. Restricted reads never use whole-table statistics.
Usage:
bijection tables analyze <collection> [flags]
Examples:
bijection tables analyze sales --fields region,amount
Flags:
--as string Act as a person, such as alice, on a development, preview or local deployment. Production refuses development identities.
--branch string Use a branch under your current business identity.
--deployment string Inspect definitions in a specific deployment. Accepts:
• a deployment name (e.g. joyful-capybara-123)
• a deployment reference (e.g. dev/james, staging)
• dev (for your personal dev deployment)
• prod (for your project’s default production deployment).
You can also select deployments in other projects with project-slug:reference.
--fields string Comma-separated fields to describe, at most 64.
-h, --help help for analyze
--help-all Show all commands and options, including administrative controls.
--json Print the exact analyzed statistics as JSON.
--url string Inspect definitions in a private deployment at this explicit origin.
