Xuanhe Zhou
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Research Interests
Xuanhe Zhou has a total of 119 co-authors. Here are the top co-authors by number of shared publications:
- Guoliang Li
Tsinghua University
20 shared publications - Wei Zhou
Shanghai Jiao Tong University, Shanghai, China
8 shared publications - Fan Wu
Shanghai Jiao Tong University, Shanghai, China
6 shared publications - Bingsheng He
National University of Singapore, Singapore, Singapore
3 shared publications - Chen Lin
School of Informatics, Xiamen University, Xiamen, China
3 shared publications - Zirui Tang
Shanghai Jiao Tong University, Shanghai, China
3 shared publications
Latest Publications
- Open Access
- 30 May 2026
- Published by Association for Computing Machinery (ACM)
- p. 580-587
Vector Search for the Future: From Memory-Resident, Static Heterogeneous Storage, to Cloud-Native Architectures
- 30 May 2026
- Published by Association for Computing Machinery (ACM)
- p. 580-587
Open AccessVector Search for the Future: From Memory-Resident, Static Heterogeneous Storage, to Cloud-Native Architectures
- Open Access
- 30 May 2026
- Published by Association for Computing Machinery (ACM)
- p. 122-125
FeatInsight: An Online ML Feature Management System on 4Paradigm Sage-Studio Platform
- 30 May 2026
- Published by Association for Computing Machinery (ACM)
- p. 122-125
Open AccessFeatInsight: An Online ML Feature Management System on 4Paradigm Sage-Studio Platform
- Open Access
- 30 May 2026
- Published by Association for Computing Machinery (ACM)
- p. 154-157
CrackSQL: A Hybrid Dialect Translation System Powered by LLM
SQL dialect translation is critical for enabling interoperability across different database systems. However, translating queries between dialects (e.g., PostgreSQL to MySQL) remains challenging due to syntactic discrepancies and subtle semantic mismatches. Existing approaches (i.e., manual translation, rule-based systems, LLM-based methods) either demand substantial engineering effort (e.g., crafting custom translation rules) or suffer from reliability issues (e.g., hallucinating non-existent functions), especially when dealing with complex queries. In this demonstration, we present CrackSQL, the first hybrid SQL dialect translation system that integrates rule-based techniques with LLM-based methods to overcome these limitations. CrackSQL is built on three key innovations: (1) A functionality-based query processing module that decomposes lengthy, complex SQLs into manageable functional units, reducing hallucination risk; (2) A cross-dialect syntax embedding model that aligns textual syntax specifications to support accurate syntax matching; (3) An adaptive local-to-global translation strategy that incrementally expands translation scope to preserve semantic correctness. CrackSQL supports three translation modes and offers flexible deployment options via a web console, PyPI package, and command-line interface, thus enabling seamless integration in both interactive and batch translation scenarios. CrackSQL is fully implemented and available at: https://github.com/weAIDB/CrackSQL.
- 30 May 2026
- Published by Association for Computing Machinery (ACM)
- p. 154-157
Open AccessCrackSQL: A Hybrid Dialect Translation System Powered by LLM
SQL dialect translation is critical for enabling interoperability across different database systems. However, translating queries between dialects (e.g., PostgreSQL to MySQL) remains challenging due to syntactic discrepancies and subtle semantic mismatches. Existing approaches (i.e., manual translation, rule-based systems, LLM-based methods) either demand substantial engineering effort (e.g., crafting custom translation rules) or suffer from reliability issues (e.g., hallucinating non-existent functions), especially when dealing with complex queries. In this demonstration, we present CrackSQL, the first hybrid SQL dialect translation system that integrates rule-based techniques with LLM-based methods to overcome these limitations. CrackSQL is built on three key innovations: (1) A functionality-based query processing module that decomposes lengthy, complex SQLs into manageable functional units, reducing hallucination risk; (2) A cross-dialect syntax embedding model that aligns textual syntax specifications to support accurate syntax matching; (3) An adaptive local-to-global translation strategy that incrementally expands translation scope to preserve semantic correctness. CrackSQL supports three translation modes and offers flexible deployment options via a web console, PyPI package, and command-line interface, thus enabling seamless integration in both interactive and batch translation scenarios. CrackSQL is fully implemented and available at: https://github.com/weAIDB/CrackSQL.
- Open Access
- 30 May 2026
- Published by Association for Computing Machinery (ACM)
- p. 320-333
ByteHouse: ByteDance's Cloud-Native Data Warehouse for Real-Time Multimodal Data Analytics
With the rapid rise of intelligent data services, modern enterprises increasingly require efficient, multimodal, and cost-effective data analytics infrastructures. However, in ByteDance's production environments, existing systems fall short due to limitations such as I/O-inefficient multimodal storage, inflexible query optimization (e.g., failing to optimize multimodal access patterns), and performance degradation caused by resource disaggregation (e.g., loss of data locality in remote storage). To address these challenges, we introduce ByteHouse (https://bytehouse.cloud), a cloud-native data warehouse designed for real-time multimodal data analytics. The storage layer integrates a unified table engine that provides a two-tier logical abstraction and physically consistent layout, SSD-backed cluster-scale cache (CrossCache) that supports shared caching across compute nodes, and virtual file system (NexusFS) that enable efficient local access on compute nodes. The compute layer supports analytical, batch, and incremental execution modes, with tailored optimizations for hybrid queries (e.g., runtime filtering over tiered vector indexes). The control layer coordinates global metadata and transactions, and features an effective optimizer enhanced by historical execution traces and AI-assisted plan selection. Evaluations on internal and standard workloads show that ByteHouse achieves significant efficiency improvement over existing systems.
- 30 May 2026
- Published by Association for Computing Machinery (ACM)
- p. 320-333
Open AccessByteHouse: ByteDance's Cloud-Native Data Warehouse for Real-Time Multimodal Data Analytics
With the rapid rise of intelligent data services, modern enterprises increasingly require efficient, multimodal, and cost-effective data analytics infrastructures. However, in ByteDance's production environments, existing systems fall short due to limitations such as I/O-inefficient multimodal storage, inflexible query optimization (e.g., failing to optimize multimodal access patterns), and performance degradation caused by resource disaggregation (e.g., loss of data locality in remote storage). To address these challenges, we introduce ByteHouse (https://bytehouse.cloud), a cloud-native data warehouse designed for real-time multimodal data analytics. The storage layer integrates a unified table engine that provides a two-tier logical abstraction and physically consistent layout, SSD-backed cluster-scale cache (CrossCache) that supports shared caching across compute nodes, and virtual file system (NexusFS) that enable efficient local access on compute nodes. The compute layer supports analytical, batch, and incremental execution modes, with tailored optimizations for hybrid queries (e.g., runtime filtering over tiered vector indexes). The control layer coordinates global metadata and transactions, and features an effective optimizer enhanced by historical execution traces and AI-assisted plan selection. Evaluations on internal and standard workloads show that ByteHouse achieves significant efficiency improvement over existing systems.
- Open Access
- 18 May 2026
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Management of Data
- Vol. 4  (3) , 1-26
Automating Database-Native Function Code Synthesis with LLMs
Database systems incorporate an ever-growing number of functions built in their kernels (a.k.a., database native functions) for scenarios like new application support and business migration. This growth causes an urgent demand for automatic database native function synthesis. While recent advances in LLM-based code generation (e.g., Claude Code) show promise, existing approaches are too generic for database-specific development. They often hallucinate or overlook critical context because database function synthesis is inherently complex and error-prone, where synthesizing a single database function may involve registering multiple function units (e.g., for different input types), placing code in the correct source files, linking internal references, and implementing logic correctly. To this end, we propose DBCooker , an LLM-based system for automatically synthesizing database native functions. The system consists of three key components. First, the function characterization module aggregates multi-source declarations, identifies function units that require specialized coding through hierarchical analysis, and traces cross-unit dependencies via static analysis. Second, we design operations to address the main synthesis challenges: (1) a pseudo-code–based coding plan generator that constructs structured implementation skeletons by identifying key elements such as reusable referenced functions; (2) a hybrid fill-in-the-blank model guided by probabilistic priors and component awareness to integrate core logic with reusable routines; and (3) three-level progressive validation , including syntax checking, standards compliance, and LLM-guided semantic verification. Finally, an adaptive orchestration strategy unifies these operations with existing database tools and dynamically sequences them based on the orchestration history of similar functions. Results show that our system outperforms state-of-the-art methods on SQLite, PostgreSQL, and DuckDB (34.55% higher accuracy on average), and can synthesize four categories of new functions absent in the latest SQLite (v3.50). The code is available at https://github.com/weAIDB/DBCooker .
- 18 May 2026
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Management of Data
- Vol. 4  (3) , 1-26
Open AccessAutomating Database-Native Function Code Synthesis with LLMs
Database systems incorporate an ever-growing number of functions built in their kernels (a.k.a., database native functions) for scenarios like new application support and business migration. This growth causes an urgent demand for automatic database native function synthesis. While recent advances in LLM-based code generation (e.g., Claude Code) show promise, existing approaches are too generic for database-specific development. They often hallucinate or overlook critical context because database function synthesis is inherently complex and error-prone, where synthesizing a single database function may involve registering multiple function units (e.g., for different input types), placing code in the correct source files, linking internal references, and implementing logic correctly. To this end, we propose DBCooker , an LLM-based system for automatically synthesizing database native functions. The system consists of three key components. First, the function characterization module aggregates multi-source declarations, identifies function units that require specialized coding through hierarchical analysis, and traces cross-unit dependencies via static analysis. Second, we design operations to address the main synthesis challenges: (1) a pseudo-code–based coding plan generator that constructs structured implementation skeletons by identifying key elements such as reusable referenced functions; (2) a hybrid fill-in-the-blank model guided by probabilistic priors and component awareness to integrate core logic with reusable routines; and (3) three-level progressive validation , including syntax checking, standards compliance, and LLM-guided semantic verification. Finally, an adaptive orchestration strategy unifies these operations with existing database tools and dynamically sequences them based on the orchestration history of similar functions. Results show that our system outperforms state-of-the-art methods on SQLite, PostgreSQL, and DuckDB (34.55% higher accuracy on average), and can synthesize four categories of new functions absent in the latest SQLite (v3.50). The code is available at https://github.com/weAIDB/DBCooker .
- Open Access
- 18 May 2026
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Management of Data
- Vol. 4  (3) , 1-26
MoDora: Tree-Based Semi-Structured Document Analysis System
Semi-structured documents integrate diverse interleaved data elements (e.g., tables, charts, hierarchical paragraphs) arranged in various and often irregular layouts. These documents are widely observed across domains and account for a large portion of real-world data. However, existing methods struggle to support natural language question answering over these documents due to three main technical challenges: (1) The elements extracted by techniques like OCR are often fragmented and stripped of their original semantic context, making them inadequate for analysis. (2) Existing approaches lack effective representations to capture hierarchical structures within documents (e.g., associating tables with nested chapter titles) and to preserve layout-specific distinctions (e.g., differentiating sidebars from main content). (3) Answering questions often requires retrieving and aligning relevant information scattered across multiple regions or pages, such as linking a descriptive paragraph to table cells located elsewhere in the document. To address these issues, we propose Modora, an LLM-powered system for semi-structured document analysis. First, we adopt a local-alignment aggregation strategy to convert OCR-parsed elements into layout-aware components, and conduct type-specific information extraction for components with hierarchical titles or non-text elements. Second, we design the Component-Correlation Tree (CCTree) to hierarchically organize components, explicitly modeling inter-component relations and layout distinctions through a bottom-up cascade summarization process. Finally, we propose a question-type-aware retrieval strategy that supports (1) layout-based grid partitioning for location-based retrieval and (2) LLM-guided pruning for semantic-based retrieval. Experiments show Modora outperforms baselines by 5.97%-61.07% in accuracy. The code is at https://github.com/weAIDB/MoDora.
- 18 May 2026
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Management of Data
- Vol. 4  (3) , 1-26
Open AccessMoDora: Tree-Based Semi-Structured Document Analysis System
Semi-structured documents integrate diverse interleaved data elements (e.g., tables, charts, hierarchical paragraphs) arranged in various and often irregular layouts. These documents are widely observed across domains and account for a large portion of real-world data. However, existing methods struggle to support natural language question answering over these documents due to three main technical challenges: (1) The elements extracted by techniques like OCR are often fragmented and stripped of their original semantic context, making them inadequate for analysis. (2) Existing approaches lack effective representations to capture hierarchical structures within documents (e.g., associating tables with nested chapter titles) and to preserve layout-specific distinctions (e.g., differentiating sidebars from main content). (3) Answering questions often requires retrieving and aligning relevant information scattered across multiple regions or pages, such as linking a descriptive paragraph to table cells located elsewhere in the document. To address these issues, we propose Modora, an LLM-powered system for semi-structured document analysis. First, we adopt a local-alignment aggregation strategy to convert OCR-parsed elements into layout-aware components, and conduct type-specific information extraction for components with hierarchical titles or non-text elements. Second, we design the Component-Correlation Tree (CCTree) to hierarchically organize components, explicitly modeling inter-component relations and layout distinctions through a bottom-up cascade summarization process. Finally, we propose a question-type-aware retrieval strategy that supports (1) layout-based grid partitioning for location-based retrieval and (2) LLM-guided pruning for semantic-based retrieval. Experiments show Modora outperforms baselines by 5.97%-61.07% in accuracy. The code is at https://github.com/weAIDB/MoDora.
- Open Access
- 4 December 2025
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Management of Data
- Vol. 3  (6) , 1-27
ST-Raptor: LLM-Powered Semi-Structured Table Question Answering
Semi-structured tables, widely used in real-world applications (e.g., financial reports, medical records, transactional orders), often involve flexible and complex layouts (e.g., hierarchical headers and merged cells). These tables generally rely on human analysts to interpret table layouts and answer relevant natural language questions, which is costly and inefficient. To automate the procedure, existing methods face significant challenges. First, methods like NL2SQL require converting semi-structured tables into structured ones, which often causes substantial information loss. Second, methods like NL2Code and multi-modal LLM QA struggle to understand the complex layouts of semi-structured tables and cannot accurately answer corresponding questions. To this end, we propose ST-Raptor, a tree-based framework for semi-structured table question answering ( semi-structured table QA ) using large language models. First, we introduce the Hierarchical Orthogonal Tree (HO-Tree), a structural model that captures complex semi-structured table layouts, along with an effective algorithm for constructing the tree by identifying headers, content values, and their implicit relationships. Second, we define a set of basic tree operations to guide LLMs in executing common QA tasks. Given a user question, ST-Raptor decomposes it into simpler sub-questions, generates corresponding tree operation pipelines, and conducts operation-table alignment for accurate pipeline execution. Third, we incorporate a two-stage verification mechanism: (1) forward validation checks the correctness of execution steps, while (2) backward validation evaluates answer reliability by reconstructing queries from predicted answers. To benchmark the performance, we present SSTQA, a dataset of 764 questions over 102 real-world semi-structured tables. Experiments show that ST-Raptor outperforms nine baselines by up to 20% in answer accuracy. The code is available at https://github.com/weAIDB/ST-Raptor.
- 4 December 2025
- Published by Association for Computing Machinery (ACM) in Proceedings of the ACM on Management of Data
- Vol. 3  (6) , 1-27
Open AccessST-Raptor: LLM-Powered Semi-Structured Table Question Answering
Semi-structured tables, widely used in real-world applications (e.g., financial reports, medical records, transactional orders), often involve flexible and complex layouts (e.g., hierarchical headers and merged cells). These tables generally rely on human analysts to interpret table layouts and answer relevant natural language questions, which is costly and inefficient. To automate the procedure, existing methods face significant challenges. First, methods like NL2SQL require converting semi-structured tables into structured ones, which often causes substantial information loss. Second, methods like NL2Code and multi-modal LLM QA struggle to understand the complex layouts of semi-structured tables and cannot accurately answer corresponding questions. To this end, we propose ST-Raptor, a tree-based framework for semi-structured table question answering ( semi-structured table QA ) using large language models. First, we introduce the Hierarchical Orthogonal Tree (HO-Tree), a structural model that captures complex semi-structured table layouts, along with an effective algorithm for constructing the tree by identifying headers, content values, and their implicit relationships. Second, we define a set of basic tree operations to guide LLMs in executing common QA tasks. Given a user question, ST-Raptor decomposes it into simpler sub-questions, generates corresponding tree operation pipelines, and conducts operation-table alignment for accurate pipeline execution. Third, we incorporate a two-stage verification mechanism: (1) forward validation checks the correctness of execution steps, while (2) backward validation evaluates answer reliability by reconstructing queries from predicted answers. To benchmark the performance, we present SSTQA, a dataset of 764 questions over 102 real-world semi-structured tables. Experiments show that ST-Raptor outperforms nine baselines by up to 20% in answer accuracy. The code is available at https://github.com/weAIDB/ST-Raptor.
- Open Access
- 3 August 2025
- Published by Association for Computing Machinery (ACM)
- p. 5960-5971
Revolutionizing Database Q&A with Large Language Models: Comprehensive Benchmark and Evaluation
- 3 August 2025
- Published by Association for Computing Machinery (ACM)
- p. 5960-5971
Open AccessRevolutionizing Database Q&A with Large Language Models: Comprehensive Benchmark and Evaluation
- Open Access
- 22 June 2025
- Published by Association for Computing Machinery (ACM)
- p. 729-742
OpenMLDB: A Real-Time Relational Data Feature Computation System for Online ML
- 22 June 2025
- Published by Association for Computing Machinery (ACM)
- p. 729-742
Open AccessOpenMLDB: A Real-Time Relational Data Feature Computation System for Online ML
- Open Access
- 22 June 2025
- Published by Association for Computing Machinery (ACM)
- p. 235-238
D-Bot: An LLM-Powered DBA Copilot
- 22 June 2025
- Published by Association for Computing Machinery (ACM)
- p. 235-238
Open AccessD-Bot: An LLM-Powered DBA Copilot