Skip to main content

Delete

delete removes rows matching a filter. Same shape as update, minus the records field.

Wire payload​

{
  "db": "mydb",
  "table": "users",
  "query": "id = 42",
  "protopass": "default"
}
FieldMeaning
db / tableTarget database and table
queryONQL filter expression selecting rows to remove
protopassProtocol identifier

query is required. An empty or missing filter is rejected — if you really want to empty a table, pass a tautology like id > 0, or use schema drop table to remove the table entirely.

Shell​

onql> delete {"db":"mydb","table":"users","query":"id = 42"}

The shell does not currently ship a shorthand for delete; the payload is sent as raw JSON. A common pattern is to first run a read to preview the rows:

onql> mydb.users[status = "archived" and updated_at < "2025-01-01"]{id, name}

...then reuse the same filter in the delete:

onql> delete {"db":"mydb","table":"users","query":"status = \"archived\" and updated_at < \"2025-01-01\""}

From a driver​

# Python
client.delete(db="mydb", table="users", query="id = 42")
// Node
await client.delete({ db: 'mydb', table: 'users', query: 'id = 42' });

Response​

Returns the rows that were removed (so you can log them, push to an audit trail, etc.), or {"error": "..."} on failure.

No soft-delete by default

delete is a hard remove. If you need soft-deletes, add a deleted_at column to your schema and update it instead; then scope reads with a context rule like shop.users[deleted_at = null]. See Concepts → Auth & Context.

Bulk delete​

Filters compose with boolean logic — there's no separate "bulk delete" call. To remove a set of rows, widen the filter:

{ "db": "mydb", "table": "sessions", "query": "expires_at < \"2026-01-01\"" }

This still runs as a single request and returns all removed rows.