Extraneous Text in SQL Query Generation with create_sql_query_chain in Langchain

[Abhishek514964]

Checked other resources

• I added a very descriptive title to this question.
• I searched the LangChain documentation with the integrated search.
• I used the GitHub search to find a similar question and didn't find it.

Commit to Help

• I commit to help with one of those options 👆

Example Code

import os
from dotenv import load_dotenv, find_dotenv
from langchain import create_sql_query_chain
from langchain.chains import QuerySQLDataBaseTool
from sqlalchemy import create_engine
from langchain_openai.chat_models import AzureChatOpenAI
from langchain.sql_database import SQLDatabase


load_dotenv(find_dotenv())

OPENAI_API_KEY = os.getenv("AZURE_OPENAI_API_KEY")
OPENAI_API_MODEL = os.getenv("AZURE_OPEN_API_MODEL")
OPENAI_API_BASE_URL = os.getenv("AZURE_OPENAI_API_BASE_URL")
OPENAI_API_VERSION = os.getenv("AZURE_OPENAI_API_VERSION")

db_path = r'C:\path\to\the\database.db'

db_url = f"sqlite:///{db_path}"
engine = create_engine(db_url)

db = SQLDatabase(engine=engine)

llm = AzureChatOpenAI(
    api_key=OPENAI_API_KEY,
    azure_endpoint=OPENAI_API_BASE_URL,
    api_version=OPENAI_API_VERSION,
    azure_deployment=OPENAI_API_MODEL,
    model="gpt-4o",  
    verbose=True
)

write_query = create_sql_query_chain(llm, db)

execute_query = QuerySQLDataBaseTool(db=db)

query_chain = (write_query | execute_query)

result = query_chain.invoke({"question": "Tell me the summary of the issues and the total time logged for assignee hdhawan, and do not use LIMIT in the query"})
    
print(result)

Description

Hello,

I am trying to create a question-answer system that interacts with an SQL database by generating SQL queries and executing them using Langchain. However, I am facing an issue when running the following code:

write_query  = create_sql_query_chain(llm, db)

execute_query = QuerySQLDataBaseTool(db=db)

query_chain = (write_query | execute_query)

print(query_chain.invoke({"question": "Tell me the summary of the issues and the total time logged for assignee hdhawan, and do not use LIMIT in the query"}))

While the query is being generated correctly, the response includes some extraneous strings such as Question and SQLQuery, as shown below:

Question: Tell me the summary of the issues and the total time logged for assignee hdhawan, and do not use LIMIT in the query
SQLQuery: 
```sql
SELECT "Summary", "Time Spent"
FROM "Zira_Issues_sql"
WHERE "Assignee" = 'hdhawan';

This results in the following error during query execution:

Error: (sqlite3.OperationalError) near "Question": syntax error
[SQL: Question: Tell me the summary of the issues and the total time logged for assignee hdhawan, and do not use LIMIT in the query
SQLQuery: 
```sql
SELECT 
    "Summary", 
    "Time Spent"
FROM 
    "Zira_Issues_sql"
WHERE 
    "Assignee" = 'hdhawan'
```]

I am expecting the chain to only generate and execute the SQL query, without the additional extraneous text (e.g., Question, SQLQuery). I followed the example from the Langchain documentation (specifically this article), but I'm still encountering this issue.

Could someone help me understand what might be causing this or suggest any steps I might be missing?

Thank you!

System Info

---

OS: Windows
OS Version: 10.0.22631
Python Version: 3.10.11 (tags/v3.10.11:7d4cc5a, Apr 5 2023, 00:38:17) [MSC v.1929 64 bit (AMD64)]