Naming Model¶
CockroachDB objects are three-part: database.schema.table. crdb-dump uses this
fully-qualified form consistently across SQL, filenames, manifests, and resume
logs.
--tables / --exclude-tables input¶
| Input | Interpreted as |
|---|---|
users |
<db>.public.users |
cpkit.tasks |
<db>.cpkit.tasks |
mydb.cpkit.tasks |
mydb.cpkit.tasks |
A two-part name means schema.table (database from --db), not the legacy
database.table. An explicit database prefix must match --db.
Identifier quoting¶
All identifiers (database, schema, table, column) are double-quoted when used in SQL, so mixed-case and reserved-word names work correctly and safely.
File and manifest naming¶
Data files and manifests use the database.schema.table stem:
mydb.public.users_001.csv
mydb.public.users.manifest.json
mydb.cpkit.tasks_001.csv
mydb.cpkit.tasks.manifest.json
Resume-log keys are derived from the same database.schema.table name.