SQL Chain Example | 🦜🔗 LangChain UseCases

SQL Chain Example | 🦜🔗 LangChain UseCases
In this video, I will walk you through a SQL Chain example. I will demonstrate how you can use the SQLDatabaseChain for answering questions over a database. I will be using SQLite and Chinook database. We will also create a database, .db using .sql file. Happy learning and chaining 😎
👉🏼 Links:
GitHub code: github.com/sud...
LangChain doc: python.langcha...
⏰ Time stamps:
00:00 👋🏼 Introduction
01:30 ⚙️ Setup
01:54 📚 Creating database in SQLite from .sql file
04:55 🤔💻 SQL Chain part
12:04 🙏 Conclusion
☕ Buy me a Coffee: ko-fi.com/data...
🔗 Other videos you might find helpful:
PrivateGPT: • PrivateGPT 🔥 | No Data...
LangChain playlist: • LangChain
Python Virtual Env: • Python Virtual Environ...
▶️ Connect with me:
*RUclips: www.youtube.co...
*LinkedIn: / sudarshan-koirala
*Twitter: / mesudarshan
#langchain #llm #openai #sql #sqlite3

[@AO-ip9hi]

Tapai ko video haru Sarai ramro cha bro! Keep up the good work.

[@datasciencebasics]

dhanyabaad :)

[@foysalmamun5106]

Many Thanks

[@praveenc2248]

Awesome, Thanks bro

[@veeru18]

how we could get answer in json format please advice

[@user-jj5et9ep7f]

Hi, thanks for the video.
I have 2 questions:
Question 1: what is the difference between "SQLDatabaseChain" and "SQLDatabaseToolkit"?
Question 2: I get an error when i try using the create_sql_agent in the below code.
toolkit = SQLDatabaseToolkit(db=db)
agent_executor = create_sql_agent(
llm=OpenAI(temperature=0),
toolkit=toolkit,
verbose=True
)
ValidationError: 1 validation error for SQLDatabaseToolkit
llm
field required (type=value_error.missing)
Can you please help here? Appretiate your help

[@rajivraghunathan9104]

Excellent info ,,, pls do one video in querying JSON Documents . No one has done any video

[@datasciencebasics]

Hello, Created one, hope you find it helpful :)

[@hadjersa28]

I made a function to compare between LLMs results and humans results. is sql Query already exist in the database or is it you who makes them ?

[@datasciencebasics]

SQL queries are created by LLMs when we ask the natural langauge.

[@python360]

Thank you, would querying a large database with say 1million row use up a lot of tokens / is there any for of caching that can be done with ChatGPT ?

[@datasciencebasics]

Hello, haven’t tested it out but you could try it out and have a look at OpenAI usage to see how it behaves.

[@Anna00722]

Can we plot a graph of thr output data

[@datasciencebasics]

this is quite old video and I haven’t tried myself yet. Give a try with latest models.

[@roberthuff3122]

Thank you. FUBAR! LOL

[@colofthedead6101]

Unfortunately, in reality this functionality is very early days and doesn't work very well, generating joins across non-existent columns and SQL that is engine specific. It's only good for playing around with at the moment.

[@datasciencebasics]

I agree, hopefully more functionality in the future as llms and usecases around it are quite in early stage.

[@colofthedead6101]

@@datasciencebasics I've persisted with this library and the more information about my database that I include in the prompt, the better the results. Less, denormalised tables help too. You can see the potential though and I've gotten some 'wow factor' from demoing it to management.

[@datasciencebasics]

@@colofthedead6101 Great findings. Eventually, everything in LLM ends in good prompt :)

[@mariocuezzo8027]

nice video! i try this with ggml-gpt4all-l13b-snoozy.bin and i have this problem : SQLQuery:The prompt size exceeds the context window size and cannot be processed.
The prompt size exceeds the context window size and cannot be processed. any ideas?