Schemas and Session Settings

KalamDB schemas group tables, types, and routines. Use PostgreSQL CREATE SCHEMA and qualify application objects so queries behave consistently across connections.

Create a KalamDB schema

1CREATE SCHEMA IF NOT EXISTS app;2CREATE TABLE app.notes (3  id BIGSERIAL PRIMARY KEY,4  body TEXT NOT NULL5) WITH (TYPE = 'USER');6SELECT * FROM app.notes;

CREATE NAMESPACE remains a compatibility alias. New schemas should use CREATE SCHEMA.

Resolve names with search_path

1SET search_path TO app, public;2SHOW search_path;3SELECT * FROM notes;4RESET search_path;

The default resolves "$user", public. A stored schema named default is distinct from public; changing the search path does not rename or merge schemas. Quoted names preserve case; unquoted names fold to lowercase.

Session settings

1SET app.request_id = 'request-123';2SELECT current_setting('app.request_id');3SELECT set_config('app.request_id', 'request-456', false);4RESET app.request_id;

A PGWire connection retains settings until changed or disconnected. HTTP SQL settings last for the request, so send SET and the statements that depend on it in the same SQL request. Do not expect a later HTTP call to inherit the previous request’s settings. For application queries, schema-qualified names avoid this dependency.

Rename and remove schemas

1ALTER SCHEMA app RENAME TO workspace;2-- Only after moving or removing its objects:3DROP SCHEMA workspace;

DROP SCHEMA ... CASCADE also removes dependent objects. Use it only when you intend to remove them.

Inspect schemas

1SELECT schema_name FROM information_schema.schemata;2SELECT nspname FROM pg_catalog.pg_namespace;

Next: Tables and constraints. Reference: PostgreSQL connections.