Factual Search on CSV, Excel and Output as Table

I am looking into possible solution to develop a SEARCH on CSV or Excel, all stored in multiple folders. Some of the CSV about 50,000 rows and some with 10 rows.

I don’t want the search to be reasoning, all I want is to be factual.

Like e.g. user input: show me apple’s products by volume between Oct 2022 to Oct 2023.
The output I expect is a Table.
Data to be fetched from CSV file_1 with columns - company, products, months, volume

user input: give me the list of patents that apple has
Output is a table
Data to be fetched from CSV file_2 with columns - company, products, patent year

How can I achieve this?

Hi and welcome to the Developer Forum!

This can already be achieved fairly trivially with the GPT-4 Advanced Data Analysis tool, if you wanted to replicate how it does this task you could start there and take a look. It essentially generates python code to load and manipulate the files and then show the results to the user.

1 Like

what are your thoughts on finchat .io? My needs are similar to what they have. Just a different use case - I have my own proprietary data that needs to be delivered to users like how finchat .io is doing it. Any advice will be greatly appreciated.

That actually sounds like a trivial SQL job. Import the CSV into a database and use chatgpt to learn SQL.