ParPaRaw: Massively Parallel Parsing of Delimiter-Separated Raw Data

Summary: GPU-parallel delimiter parsing via FSM simulation eliminates the sequential context pass, supporting expressive formats such as quoted fields. Scales to thousands of cores at 14.2 GB/s, with streaming PCIe transfers enabling 4.8 GB end-to-end parsing in 0.44 s. (summarized by gpt-5.6-luna on Jul 24 2026)

Paper ID: h7bdfeed1648b01c8
Venue: VLDB
Year: 2020
Pagerank: 5.8759961e-05
Overall Rank: 6,079 | 59.26%
DOI: 10.14778/3377369.3377372
PDF: Download (CC BY-NC-ND 4.0)

Incoming Non-self Citations Over Time

Authors

1. Elias Stehle (Technical University of Munich)
2. Hans-Arno Jacobsen (Technical University of Munich)

BibTeX Citation

@article{stehle_vldb20,
        title = {{ParPaRaw: Massively Parallel Parsing of Delimiter-Separated Raw Data}},
        author = {Stehle, Elias and Jacobsen, Hans-Arno},
        journal = {PVLDB},
        series = {{VLDB} '20},
        volume = {13},
        number = {5},
        pages = {616--628},
        doi = {10.14778/3377369.3377372},
        url = {https://doi.org/10.14778/3377369.3377372},
        year = {2020}
}

Incoming Citations (Sorted by Pagerank)

Showing 7 of 7 citing papers.

Rank	Citing Paper	Year	Venue	Pagerank
6,139	Parallel Index-based Stream Join on a Multicore CPU	2020	SIGMOD	5.8567401e-05
7,945	Scalable Structural Index Construction for JSON Analytics	2021	VLDB	5.4051378e-05
9,571	GIO: Generating Efficient Matrix and Frame Readers for Custom Data Formats by Example	2023	SIGMOD	5.1430298e-05
10,176	GpJSON: High-performance JSON Data Processing on GPUs	2025	VLDB	5.0557339e-05
10,340	Fast and Scalable Data Transfer Across Data Systems	2025	SIGMOD	5.0210286e-05
10,362	Distributed Stream KNN Join	2021	SIGMOD	5.014328e-05
10,951	One Pass to Parse Them All: Fused Parallel CSV Processing	2026	VLDB	4.9605343e-05

Outgoing Citations (Sorted by Pagerank)

Showing 17 of 17 cited papers.

Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.

Rank	Cited Paper	Year	Venue	Pagerank
1,072	NoDB: Efficient Query Execution on Raw Data Files	2012	SIGMOD	0.00012143482
1,937	Here are my Data Files. Here are my Queries. Where are my Results?	2011	CIDR	9.3333463e-05
1,942	Instant Loading for Main Memory Databases	2013	VLDB	9.3133693e-05
2,342	Parallel In-Situ Data Processing with Speculative Loading	2014	SIGMOD	8.5978326e-05
2,356	Filter Before You Parse: Faster Analytics on Raw Data with Sparser	2018	VLDB	8.5715425e-05
2,437	Parallel Data Analysis Directly on Scientific File Formats	2014	SIGMOD	8.4576748e-05
2,446	Mison: A Fast JSON Parser for Data Analytics	2017	VLDB	8.4392802e-05
2,485	A Memory Bandwidth-Efficient Hybrid Radix Sort on GPUs	2017	SIGMOD	8.3800709e-05
2,741	Speculative Distributed CSV Data Parsing for Big Data Analytics	2019	SIGMOD	8.0574455e-05
2,936	Constance: An Intelligent Data Lake System	2016	SIGMOD	7.825556e-05
3,079	Adaptive Query Processing on RAW Data	2014	VLDB	7.6544931e-05
3,383	A Hybrid B+-tree as Solution for In-Memory Indexing on CPU-GPU Heterogeneous Computing Platforms	2016	SIGMOD	7.3393628e-05
3,510	Fast Queries Over Heterogeneous Data Through Engine Customization	2016	VLDB	7.2263115e-05
4,142	The bionic DBMS is coming, but what will it look like?	2013	CIDR	6.7684359e-05
4,855	ReCache: Reactive Caching for Fast Analytics over Heterogeneous Data	2018	VLDB	6.3657929e-05
5,166	FAD.js: Fast JSON Data Access Using JIT-based Speculative Optimizations	2017	VLDB	6.2338839e-05
7,298	FishStore: Fast Ingestion and Indexing of Raw Data	2019	VLDB	5.5467355e-05

Semantically Similar Papers

#	Overall Rank	Paper	Year	Venue
1	8,095	Parallel k-Core Decomposition: Theory and Practice	2025	SIGMOD
2	9,695	Parallel Query Processing: To Separate Communication from Computation	2022	SIGMOD
3	3,079	Adaptive Query Processing on RAW Data	2014	VLDB
4	6,553	A Holistic View of Stream Partitioning Costs	2017	VLDB
5	3,657	Realtime Top-k Personalized PageRank over Large Graphs on GPUs	2020	VLDB
6	5,039	Massive Scale-out of Expensive Continuous Queries	2011	VLDB
7	2,342	Parallel In-Situ Data Processing with Speculative Loading	2014	SIGMOD
8	2,356	Filter Before You Parse: Faster Analytics on Raw Data with Sparser	2018	VLDB
9	2,741	Speculative Distributed CSV Data Parsing for Big Data Analytics	2019	SIGMOD
10	10,951	One Pass to Parse Them All: Fused Parallel CSV Processing	2026	VLDB