Prompt as a Data Type: In-Database LLM Prompt Management and Rewriting

2026-07-23

AI Summary

This work addresses the limitation that large language model (LLM) prompts are typically stored externally in unstructured forms, rendering them invisible to query optimization and metadata management within database systems. To overcome this, the paper introduces PROMPT as a first-class logical data type, enabling prompts—along with their templates, attribute bindings, and metadata—to be stored directly in tables or views. It further proposes an EVAL operator to render and execute these prompts. Leveraging query optimizer principles and reflective programming, the system dynamically rewrites prompts using database metadata, thereby opening a new dimension for optimization. Experimental results demonstrate that this approach significantly improves output validity on both synthetic and real-world datasets while achieving a better trade-off between cost and quality.

Abstract

Large Language Models (LLMs) are increasingly used in database-backed applications to classify tuples, filter records using semantic predicates, extract structured attributes, and enrich query results. Yet the prompt that start these computations are typically stored outside the DBMS in unstructured formats, making them invisible to query execution, metadata management, and optimization. Drawing on Stonebraker's QUEL as a Data Type and the principles of reflective programming, this paper introduces PromptDB, a database system that treats prompts as tuple-level database values. PromptDB provides a logical PROMPT datatype whose values store a template, bindings to tuple attributes, model metadata, and task metadata. Relations may contain PROMPT attributes directly in base tables, or expose them through views over joined tuples. Users query prompt-valued attributes through generated evaluation views, while the system internally renders, rewrites, optimizes, and executes prompts through an EVAL operator. Making prompts database-visible creates a new optimization space. The key idea is to bring query-optimizer thinking to prompts: just as query optimizers exploit database metadata to rewrite SQL plans, PromptDB exploits database metadata to rewrite prompts. We evaluate PromptDB on synthetic and real-world data workloads across different tasks. The results show how database-guided rewriting improves output validity and yields favorable cost-quality trade-offs compared with static, manually written prompts.

Problem

• Large Language Models
• Prompt Management
• Database Optimization
• Semantic Querying
• Metadata Visibility

Innovation

• Prompt as Data Type
• In-Database Prompt Management
• Prompt Rewriting
• LLM Integration with DBMS
• Reflective Programming