Simple SELECT statement on existing table with SQLAlchemy

Question

Nowhere on the internet does there exist a simple few-line tutorial on a simple SELECT statement for SQLAlchemy 1.0.

Assuming I've established my database connection using create_engine(), and my database tables already exist, I'd like to know how to execute the following query:

select
    name,
    age
from
    users
where
    name = 'joe'
    and
    age = 100

Answer 1

Found this while trying to figure out the same thing.

To select data from a table via SQLAlchemy, you need to build a representation of that table within SQLAlchemy. If Jupyter Notebook's response speed is any indication, that representation isn't filled in (with data from your existing database) until the query is executed.

You need Table to build a table. You need select to select data from the database. You need metadata... for reasons that aren't clear, even in the docs.

from sqlalchemy import create_engine, select, MetaData, Table, and_

engine = create_engine("dburl://user:pass@database/schema")
metadata = MetaData(bind=None)
table = Table(
    'table_name', 
    metadata, 
    autoload=True, 
    autoload_with=engine
)

stmt = select([
    table.columns.column1,
    table.columns.column2
]).where(and_(
    table.columns.column1 == 'filter1',
    table.columns.column2 == 'filter2'
))

connection = engine.connect()
results = connection.execute(stmt).fetchall()

You can then iterate over the results. See SQLAlchemy query to return only n results? on how to return one or only a few rows of data, which is useful for slower/larger queries.

for result in results:
    print(result)

I checked this with a local database, and the SQLAlchemy results are not equal to the raw SQL results. The difference, for my data set, was in how the numbers were formatted. SQL returned float64 (e.g., 633.07), while SQLAlchemy returned objects (I think Decimal, e.g. 633.0700000000.)

Some help from DataCamp's Introduction to Databases in Python

Answer 2

Since the original question has two columns in the select statement, and it can confuse some people on how to write using that:

from sqlalchemy import and_  
stmt = select([users.columns.name,users.columns.age])  
stmt= stmt.where(and_(name=='joe',age==100)  
for res in connection.execute(stmt):  
    print(res)

Answer 3

Sticking with SQL alchemy for this seems overcomplicated. What you can do instead, is to pass SQL alchemy engine component, to pandas.read_sql(sql,conn). https://pandas.pydata.org/docs/reference/api/pandas.read_sql.html

from sqlalchemy import create_engine
import pandas as pd

engine = create_engine(.....)
sql = "select name, age from users where name = 'joe' and age = 100"
df = pd.read_sql(sql,con=engine)