SQL Alchemy Parametrized Query , binding table name as parameter gives error

Question

I am using parametrized query utilizing Text object in SQL alchemy and are getting different result.

Working example:

import sqlalchemy as sqlal
from sqlalchemy.sql import text

    db_table = 'Cars'
    id_cars = 8
    query = text("""SELECT * 
                    FROM Cars 
                    WHERE idCars = :p2
                 """)
    self.engine.execute(query, {'p2': id_cars})

Example that produces sqlalchemy.exc.ProgrammingError : ( pymysql.err.ProgrammingError ) (1064, "You have an error in your SQL syntax)

import sqlalchemy as sqlal
from sqlalchemy.sql import text

    db_table = 'Cars'
    id_cars = 8
    query = text("""SELECT * 
                    FROM :p1 
                    WHERE idCars = :p2
                 """)
    self.engine.execute(query, {'p1': db_table, 'p2': id_cars})

Any idea on how I can run the query with a dynamic table name that are also protected from sql injection?

Answer 1

I use PostgreSQL and psycopg2 backend. I was able to do it using:

from psycopg2 import sql
from sqlalchemy import engine

connection: sqlalchemy.engine.Connection
connection.connection.cursor().execute(
    sql.SQL('SELECT * FROM {} where idCars = %s').format(sql.Identifier(db_table)),
    (id_cars, )
)

Answer 2

You could just use the benefits of writing in python:

Library to use:

import sqlalchemy
from sqlalchemy import create_engine, MetaData, Table, func, event
from sqlalchemy.sql import text
from urllib.parse import quote_plus

connection (that I did not see in your example - here connection with sql azure):

params = urllib.parse.quote_plus(r'...')
conn_str = 'mssql+pyodbc:///?odbc_connect={}'.format(params)
engine_azure = create_engine(conn_str, echo=True)

Your example:

db_table = 'Cars'
id_cars = 8
query = text('SELECT * FROM ' + db_table + 'WHERE idCars = ' + id_cars)
connection = engine.connect()
connection.execute(query)
connection.close()