Late Materialization (LM) is a critical technique applied in traditional column stores to speed up analytical queries. However, with modern analytical databases evolved to incorporate a vectorized columnar execution engine, LM's benefits in I/O reduction and fast columnar query processing have diminished. In this paper, we redefine the concept of Late Materialization in the context of modern analytical databases and propose Selective Late Materialization (SLM) to allow each attribute in a query to choose its own materialization point that yields the minimum cost. SLM expands the solution space of the traditional materialization problem from one unified hard-coded binary decision (i.e., early or late) for all attributes to per attribute per query decisions. By integrating SLM into DuckDB, we show that SLM consistently outperforms the baselines of Early Materialization and Late Materialization by 14.7% and 8.9%, respectively, on average using the Join Order Benchmark (JOB), with up to 76.7% latency reduction for individual queries. We observe similar results for the TPC-DS benchmark.

The use of Retrieval-Augmented Generation (RAG) has improved Large Language Models (LLMs) in collaborating with external data, yet significant challenges exist in real-world scenarios. In areas such as academic literature and finance question answering, data are often found in raw text and tables in HTML or PDF formats, which can be lengthy and highly unstructured. In this paper, we introduce a benchmark suite, namely Unstructured Document Analysis (UDA), that involves 2,965 real-world documents and 29,590 expert-annotated Q&A pairs. We revisit popular LLM- and RAG-based solutions for document analysis and evaluate the design choices and answer qualities across multiple document domains and diverse query types. Our evaluation yields interesting findings and highlights the importance of data parsing and retrieval. We hope our benchmark can shed light and better serve real-world document analysis applications. The benchmark suite and code can be found at https://github.com/qinchuanhui/UDA-Benchmark

Columnar storage is a core component of a modern data analytics system. Although many database management systems (DBMSs) have proprietary storage formats, most provide extensive support to open-source storage formats such as Parquet and ORC to facilitate cross-platform data sharing. But these formats were developed over a decade ago, in the early 2010s, for the Hadoop ecosystem. Since then, both the hardware and workload landscapes have changed. In this paper, we revisit the most widely adopted open-source columnar storage formats (Parquet and ORC) with a deep dive into their internals. We designed a benchmark to stress-test the formats' performance and space efficiency under different workload configurations. From our comprehensive evaluation of Parquet and ORC, we identify design decisions advantageous with modern hardware and real-world data distributions. These include using dictionary encoding by default, favoring decoding speed over compression ratio for integer encoding algorithms, making block compression optional, and embedding finer-grained auxiliary data structures. We also point out the inefficiencies in the format designs when handling common machine learning workloads and using GPUs for decoding. Our analysis identified important considerations that may guide future formats to better fit modern technology trends.