Feature Description

Problem Statement

Current reliance on generic SQL/DBAPI tools for connecting LlamaIndex to high-performance distributed query engines like Trino introduces critical production instability through resource exhausting, missing performance context and inefficient RAG input.

Reason

Currently LlamaIndex doesn't have a datareader purpose built for loading data from Trino so users have to hack generic tools together

Value of Feature

Proposed Solution:

We propose a native TrinoReader built around the core trino-python-client library. This olution provides a production-grade interface by focusing on rliability and RAG optimization. This will enable Gauranteed Resource Safety, Rag-Optimized Output and a clean trino-native API interface for Trino developers