SQLAlchemy Core - Multiple Table Delete

Last Updated : 14 Feb, 2022

In this article, we are going to see how to perform multiple-table DELETE in SQLAlchemy against a PostgreSQL database in Python.

Creating table for demonstration - BOOKS

Import necessary functions from the SQLAlchemy package. Establish connection with the PostgreSQL database using create_engine() function as shown below, create a table called books with columns book_id and book_price, genre, book_name. Insert record into the tables using insert() and values() function as shown.

Python3

# import necessary packages
import sqlalchemy
from sqlalchemy import create_engine, MetaData, 
Table, Column, Numeric, Integer, VARCHAR, update, text, delete
from sqlalchemy.engine import result

# establish connections
engine = create_engine(
    "database+dialect://username:password@hostname:port/databasename")

# initialize the Metadata Object
meta = MetaData(bind=engine)
MetaData.reflect(meta)

# create a table schema
books = Table(
    'books', meta,
    Column('book_id', Integer, primary_key=True),
    Column('book_price', Numeric),
    Column('genre', VARCHAR),
    Column('book_name', VARCHAR)
)

meta.create_all(engine)

# insert records into the table
statement1 = books.insert().values(book_id=1,
                                   book_price=12.2,
                                   genre='fiction',
                                   book_name='Old age')
statement2 = books.insert().values(book_id=2, 
                                   book_price=13.2,
                                   genre='non-fiction',
                                   book_name='Saturn rings')
statement3 = books.insert().values(book_id=3,
                                   book_price=121.6,
                                   genre='fiction', 
                                   book_name='Supernova')
statement4 = books.insert().values(book_id=4,
                                   book_price=100,
                                   genre='non-fiction',
                                   book_name='History of the world')
statement5 = books.insert().values(book_id=5, 
                                   book_price=1112.2,
                                   genre='fiction',
                                   book_name='Sun city')

# execute the insert records statement
engine.execute(statement1)
engine.execute(statement2)
engine.execute(statement3)
engine.execute(statement4)
engine.execute(statement5)

# Get the `books` table from the Metadata object
BOOKS = meta.tables['books']

Output:

Creating table for demonstration - book_publisher

Import necessary functions from the SQLAlchemy package. Establish connection with the PostgreSQL database using create_engine() function as shown below, create a table called book_publisher with columns publisher_id, publisher_name, and publisher_estd. Insert record into the tables using insert() and values() function as shown.

Python3

# import necessary packages
import sqlalchemy
from sqlalchemy import create_engine, MetaData,
Table, String, Column, Numeric, Integer, VARCHAR,
update, text, delete
from sqlalchemy.engine import result

# establish connection
engine = create_engine(
    "database+dialect://username:password@hostname:port/databasename")

# store engine objects
meta = MetaData()

# create a table
book_publisher = Table(
    'book_publisher', meta,
    Column('publisher_id', Integer, 
           primary_key=True),
    Column('publisher_name', String),
    Column('publisher_estd', Integer),
)

# use create_all() function to create
# a table using objects stored in meta.
meta.create_all(engine)

# insert values
statement1 = book_publisher.insert().values(
    publisher_id=1, publisher_name="Oxford",
  publisher_estd=1900)
statement2 = book_publisher.insert().values(
    publisher_id=2, publisher_name='Stanford',
  publisher_estd=1910)
statement3 = book_publisher.insert().values(
    publisher_id=3, publisher_name="MIT",
  publisher_estd=1920)
statement4 = book_publisher.insert().values(
    publisher_id=4, publisher_name="Springer",
  publisher_estd=1930)
statement5 = book_publisher.insert().values(
    publisher_id=5, publisher_name="Packt",
  publisher_estd=1940)

engine.execute(statement1)
engine.execute(statement2)
engine.execute(statement3)
engine.execute(statement4)
engine.execute(statement5)

# Get the `book_publisher` table from the Metadata object
book_publisher = meta.tables['book_publisher']

Output:

Implementing a query to perform multiple-table delete in SQLAlchemy

Performing multiple-table delete has a slightly different procedure than that of a conventional SQL query which is shown below

from sqlalchemy import delete

delete(tablename_1).where(tablename_1.c.column_name== tablename_2.c.column_name).where(tablename_2.c.column_name== 'value')

Get the books and book_publisher table from the Metadata object initialized while connecting to the database. Pass the delete query to the execute() function and get all the results using fetchall() function. Use a for loop to iterate through the results.

The SQLAlchemy query shown in the below code deletes the record with book_id in books table corresponding with publisher_name "Springer" in the book_publisher table. . Then, we can write a conventional SQL query and use fetchall() to print the results to check whether the table record is deleted properly.

Python3

from sqlalchemy import delete

# query to multiple table delete
delete_stmt = (delete(BOOKS).where(
  BOOKS.c.book_id == book_publisher.c.publisher_id).where(
    book_publisher.c.publisher_name == 'Springer'))

# execute the statement
engine.execute(delete_stmt)

# write the SQL query inside the
# text() block to fetch all records
sql = text("SELECT * from BOOKS")

# Fetch all the records
result = engine.execute(sql).fetchall()

# View the records
for record in result:
    print("\n", record)