The community member is new to the community and has a question about how LlamaIndex can connect LLM with SQL. They have read articles claiming that LlamaIndex can only query one table with high accuracy and fails when there are more than one table. The comments suggest that LlamaIndex can work with multiple tables, but for over 10 tables, a table retriever may be needed to trim down the query space. It also helps to provide the tables with extra context information. The comments indicate that the number of rows in a table does not matter, and the LLM can generate a SQL query to find customers that ordered with a value of $1000 or more and the order must contain a discount.
Hello, I am new to this community. I have a question on how LlamaIndex can connect LLM with SQL. If you have complex tables and you are trying to run join queries for example, is that something LlamaIndex can support?
It's worked fine for me in the past with multiple tables, But if you have over, say, 10 tables, you'll likely want to use some kind of table retriever to trim down the query space
How effective is it when you have a lot of data in all these tables. Let's say we have 100,000 customers with 1M orders total. Can it query something like this:
Find the customers that ordered with value $1000 or more and the order must contain a discount?
