Standard library

nox.orm

A micro-ORM: describe a table once, then create it and run parameterised insert/update/delete/select helpers over any DbConnection. It is a query builder with typed values, not an object mapper — there is no reflection in Nox, so results come back as Rows and you convert them to your own types.

Text

from nox.orm import Table, Column, Value
import nox.orm

Capability: none (it works over whichever driver you pass).

Describing a table#

Constructor	Description
`Column(name, kind, is_primary_key)`	a column; `kind` is `0` integer, `1` text, `2` real (float)
`Table(name, columns)`	a table with its columns

Values#

SQL values are wrapped in Values so every statement is parameterised:

Function	Description
`val_str(s)`, `val_int(i)`, `val_float(f)`, `val_bool(b)`, `val_null()`	construct a `Value`

Operations#

Function	Description
`create_table_sql(table)`	the `CREATE TABLE IF NOT EXISTS …` text
`create_table(conn, table)`	runs it
`insert(conn, table, values)`	inserts one row from a `dict[str, Value]` (column → value); returns the affected row count
`update(conn, table, values, where_sql, where_params)`	sets the given columns on rows matching `where_sql` (using `?` placeholders filled from `where_params`); returns the affected row count
`delete(conn, table, where_sql, where_params)`	deletes matching rows; returns the affected row count
`select(conn, table, where_sql, where_params)`	returns the matching rows as `list[Row]` in column order

where_sql is a fragment such as "age > ? AND name = ?"; its values always come from where_params. Pass "" for no condition. Column and table names are placed into the SQL text, so they must come from your code, never from user input. Errors raise OrmError (or the driver's error).

Nox

import nox.sqlite
import nox.orm
from nox.sqlite import Connection
from nox.db import Row
from nox.orm import Table, Column, Value

conn: Connection = nox.sqlite.open(":memory:")
people: Table = Table("people", [Column("id", 0, True), Column("name", 1, False), Column("age", 0, False)])
print(nox.orm.create_table_sql(people))
nox.orm.create_table(conn, people)

vals: dict[str, Value] = {"name": nox.orm.val_str("grace"), "age": nox.orm.val_int(85)}
print(nox.orm.insert(conn, people, vals))
print(nox.orm.update(conn, people, {"age": nox.orm.val_int(86)}, "name = ?", [nox.orm.val_str("grace")]))
found: list[Row] = nox.orm.select(conn, people, "age > ?", [nox.orm.val_int(50)])
print(found[0].get_str(1), found[0].get_int(2))
print(nox.orm.delete(conn, people, "name = ?", [nox.orm.val_str("grace")]))
conn.close()

Output

CREATE TABLE IF NOT EXISTS people (id INTEGER PRIMARY KEY, name TEXT, age INTEGER)
1
1
grace 86
1

Notes#

The SQL it generates uses ? placeholders; with PostgreSQL use Statement directly, or a driver that accepts ? (SQLite, MySQL).
For anything beyond single-table CRUD (joins, aggregates) write the SQL yourself with query and prepare.

Edit this page on GitHub