Seed a disposable Postgres for integration tests
Starting a Postgres container in CI takes seconds. Filling it with data that is referentially intact and identical on every run is the part that decides whether your integration suite proves anything.
Technical articles, guides, and updates from the Weavori team.
Starting a Postgres container in CI takes seconds. Filling it with data that is referentially intact and identical on every run is the part that decides whether your integration suite proves anything.
A generated row can satisfy every constraint and still be nonsense. Closing that gap takes four controls, not one: column distributions, null rates, a bounded time window, and ordering between related rows.
An honest roundup of PostgreSQL test data tools: schema-driven generators, in-code libraries, anonymizers, and enterprise platforms, with pricing and where each one wins.
Postgres data masking and subsetting are two halves of one job. Copy a small, connected slice of production into staging with the PII replaced, the keys left intact, and the substitutions stable enough to join on.
PocketOS's backups died with the volume they protected. The fix is not better backups. Split your tables by derivability and give the recipe-class half a rebuild command instead of a copy.
A nine-second agent call deleted PocketOS's database and its backups, because the backups lived inside the volume. Sandboxes isolate the machine; the credential is the boundary. Make the database disposable.
ADD CONSTRAINT UNIQUE runs in milliseconds and locks nothing that matters, then aborts on a duplicate email. Fixtures hide it and so does uniform random data. Here is the data shape that exposes it.
Your CI database picks different plans than production because its statistics differ. Inject production stats with PostgreSQL 18, or generate rows that carry the distribution. Here is how to tell which one you need.
Paste your CREATE TABLE statements and generate referentially intact Postgres data. No database, no connection string, no config. Weavori's DDL paste mode runs from a schema file.
Connect Weavori's MCP server to Claude or Cursor and your AI assistant can introspect a Postgres schema, estimate a dataset, and generate FK-intact synthetic data, in natural language.
Weavori is live: a zero-config CLI that turns any PostgreSQL schema into realistic, referentially intact synthetic data in one command. Free tier, flat-fee Pro, 14-day trial, no credit card.
Referential integrity is the floor, not the goal. Realistic data also needs production's column distributions, its null rates, and consistent ordering between rows in different tables. Here is the order to fix them in.