litellm/docs/extras/use_cases/tabular/sql_query.ipynb

SQL Query

This notebook walks through how to load and run a chain that constructs SQL queries that can be run against your database to answer a question. Note that this ONLY constructs the query and does not run it.

In [4]:

from langchain.chains import create_sql_query_chain

from langchain.chat_models import ChatOpenAI
from langchain.utilities import SQLDatabase

In [7]:

db = SQLDatabase.from_uri("sqlite:///../../../../notebooks/Chinook.db")

In [8]:

chain = create_sql_query_chain(ChatOpenAI(temperature=0), db)

In [12]:

response = chain.invoke({"question":"How many employees are there"})

In [14]:

print(response)

SELECT COUNT(*) FROM Employee

In [15]:

db.run(response)

Out [15]:

'[(8,)]'

In [ ]: