Search arXiv⌕ Search
Home/Archive/arXiv

arXiv · 2605.08687

PrepBench: How Far Are We from Natural-Language-Driven Data Preparation?

Jingzhe Xu·Rui Wang·Jiannan Wang·Guoliang Li
View original source ↗Read source document ↗Export metadata ↓

Abstract

Data preparation is a central and time-consuming stage in data analysis workflows. Traditionally, commercial tools have relied on graphical user interfaces (GUIs) to simplify data preparation, allowing users to define transformations through visual operators and workflows. Recent advances in large language models (LLMs) raise the possibility of a paradigm shift toward natural language (NL)-driven data preparation, in which users can specify preparation intents in NL directly. However, it remains unclear how far current LLM-based agents are from this paradigm shift in practice. Existing code generation benchmarks do not capture key characteristics of data preparation, including ambiguous user intents, imperfect real-world data, and the need to translate code into interpretable workflows for validation. To bridge this gap, we present PrepBench, a benchmark designed to evaluate NL-driven data preparation along three core capabilities: interactive disambiguation, prep-code generation, and code-to-workflow translation. We crawl data from the Preppin' Data Challenges, and then extend it into a systematically designed benchmark. The benchmark covers diverse domains, and each task involves 3 to 18 data preparation steps. Nearly half of the tasks require over 100 lines of Python code, and the longest solutions approach 300 lines. Our evaluation shows that, despite recent progress, realizing this paradigm shift remains challenging for state-of-the-art LLMs. PrepBench provides a principled benchmark for measuring this gap and helps identify key challenges toward realizing NL-driven data preparation.

Explore related subjects

cs.DBcs.AI

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Jingzhe Xu, Rui Wang, Jiannan Wang, Guoliang Li. 2026-05-09. PrepBench: How Far Are We from Natural-Language-Driven Data Preparation?. https://arxiv.org/abs/2605.08687

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

VectorMaton: Efficient Vector Search with Pattern Constraints via an Enhanced Suffix Automaton

Approximate nearest neighbor search (ANNS) has become a cornerstone in modern vector database systems. Given a query vector, ANNS retrieves the closest vectors from a set of base vectors. In real-world applications, vectors are often accompanied by additional information, such as sequences or structured attributes, motivating the need for fine-grained vector search with constraints on this auxiliary data. Existing methods support attribute-based filtering or range-based filtering on categorical and numerical attributes, but they do not support pattern predicates over sequence attributes. In relational databases, predicates such as LIKE and CONTAINS are fundamental operators for filtering records based on substring patterns. As vector databases increasingly adopt SQL-style query interfaces, enabling pattern predicates over sequence attributes (e.g., texts and biological sequences) alongside vector similarity search becomes essential. In this paper, we formulate a novel problem: given a set of vectors each associated with a sequence, retrieve the nearest vectors whose sequences contain a given query pattern. To address this challenge, we propose VectorMaton, an automaton-based index that integrates pattern filtering with efficient vector search, while maintaining an index size comparable to the dataset size. Extensive experiments on real-world datasets demonstrate that VectorMaton consistently outperforms all baselines, achieving up to 10x higher query throughput at the same accuracy and up to 18x reduction in index size.

cs.DB↗
arXiv2026

Efficient K-generalizable Learned Search

Learned top-K search improves the accuracy-latency trade-off of graph-based vector search, but existing methods are designed for a fixed K: serving production workloads with varying K values requires preprocessing cost proportional to the number of distinct Ks served - prohibitive in practice. This paper shows that learned search can support arbitrary K with the preprocessing cost of a single top-1 model. The key idea is to reduce top-K learned search to repeated masked top-1 refinement, which works because the distance-reduction trajectory for discovering the next top-1 vector is largely invariant to the number of results already found. We therefore train the model on trajectory features that remain effective under masking. To make repeated refinement robust and efficient, OMEGA counters error accumulation across iterations with rank-wise confidence allocation, and skips unnecessary model invocations with a statistical forecast of recall from partial results. Across nine dataset-scale configurations, OMEGA meets the 0.95 recall target with one K-independent model. Under the lowest-preprocessing configuration of each learned baseline,it reduces mean latency by 7-36% versus DARTH, 3-25% versus MultiK-DARTH, and 8-21% versus LAET on BIGANN, BIGANN-1B, DEEP, and three production workloads. On GIST, Text2Image, and MS MARCO, its latency remains within 9% of DARTH and MultiK-DARTH. On production traces, OMEGA further reduces total serving and preprocessing computation by up to 28%.

cs.DB↗
arXiv2026

Samyama: A Unified Graph-Vector Database with In-Database Optimization, Agentic Enrichment, and Hardware Acceleration

Modern data architectures fragment across graph databases, vector stores, analytics engines and optimization solvers, forcing ETL between them. We present Samyama, a graph-vector database in Rust that unifies these workloads in one engine: a RocksDB-backed store with MVCC, a vectorized executor, a cost-based planner, a CSR analytics engine, RDF and SPARQL, 22 metaheuristic solvers callable from the query language, HNSW vector indexing, and agentic enrichment that expands a graph via LLMs. It has been run to billion-edge scale: 74.3M nodes and 1.07B edges from four biomedical sources on one machine for $2.50 of spot compute. This version re-measures the system at release v1.8.0 on rented Linux hosts a reader can boot, replacing earlier figures taken on a Mac Mini that nobody could re-run. On a 16-vCPU cloud instance, ingestion reaches 123K-152K nodes/s and PageRank costs 6.7 ms per iteration at ten thousand nodes and 236 ms at a million. The previously reported 8.2x GPU speedup is withdrawn; in its place we measure the CUDA path, which no earlier version measured at all, on a rented NVIDIA A16: parity with the CPU at ten thousand nodes, 1.62x at a hundred thousand and 2.70x at a million, with 1.1-1.35 s of context initialisation on the first call. The earlier Cypher-throughput figure, its claim of near-constant index-driven scaling, and a vector-search figure are withdrawn as unreproducible. openCypher conformance, once estimated at ~90%, is now measured at 99.9% of evaluated TCK scenarios. Against Neo4j 5 and FalkorDB on one host over an identical SF10 extract, 14 of 14 SNB complex reads land within 5x of the best competitor and 7 of 14 are faster.

cs.DB↗