2 views collected around this technical thread.
DuckDB's DuckLake is an open‑standard, SQL‑driven data lake and catalog format that simplifies lakehouse architecture by managing metadata in a database while storing data in scalable Parquet files, offering multi‑user collaboration, time‑travel queries, and MIT licensing.
The article introduces Archer, a new big‑data warehouse engine built on Iceberg that adds an inverted‑index mechanism using Tantivy to provide full‑text and JSON search, storage‑compute separation, and significant performance gains over traditional Elasticsearch and Iceberg connectors.
The article proposes replacing the traditional multi‑layered data‑warehouse architecture (ODS‑DWD‑DWS‑ADS) with a single, column‑store wide‑table per business theme, achieving roughly 30 % storage savings and faster queries, while acknowledging higher ETL complexity, back‑tracking costs, and production timing challenges.
This article provides a comprehensive introduction to Apache Parquet, covering its origins, columnar storage advantages, nested schema support, internal architecture, storage model components, comparison with ORC, and practical tools for inspecting Parquet files.
This article introduces Apache Iceberg as an open table format for massive analytic datasets, explains its underlying concepts such as schema, partitioning, statistics, and read/write APIs, compares it with Hive Metastore, outlines its ACID commit process, highlights the performance and operational advantages for big‑data workloads, and previews upcoming community features.
Apache Beam 2.28.0 introduces extensive Parquet support, new hash functions in BeamSQL and ZetaSQL, ApproximateDistinct via HLL, enhanced I/O connectors including SpannerIO for Numeric fields, ParquetIO schema support, KafkaTableProvider thrift, HadoopFormatIO key/value cloning skip, and various other improvements.
Lyft transformed its legacy data pipeline by designing a cloud‑native, Flink‑based near real‑time analytics platform that ingests billions of events, writes Parquet files to S3, leverages Presto for interactive queries, and implements multi‑stage non‑blocking ETL, fault‑tolerant back‑fill, and extensive performance optimizations.
This article examines how to run several Spark jobs concurrently on a shared SparkContext, balancing full CPU‑vcore utilization with the need to generate fewer Parquet files, and presents practical experiments, scheduling strategies, and performance results.
Apache Parquet is a language‑agnostic, columnar storage format for the Hadoop ecosystem that offers high compression, efficient I/O through column and predicate push‑down, nested‑structure support, and a three‑layer architecture, and is compared with ORC while providing tooling for schema inspection.
The article introduces Kyligence's Kylin on Parquet solution, explains its plug‑in architecture, reasons for replacing HBase with Parquet, details the new Spark‑based build and query engines, auto‑tuning, global dictionary, fault‑tolerance features, and presents performance comparisons with Kylin 3.0.
This article provides a comprehensive overview of Apache Parquet, covering its purpose, architectural components, nested data model, file structure, practical Hive commands for creating and inspecting Parquet tables, and a brief introduction to the TPC‑DS benchmark for performance testing.
This article reviews major big‑data storage and query architectures—including HBase, Dremel/Parquet, pre‑aggregation systems, Lucene, and the custom Tindex solution—evaluating their strengths, weaknesses, and suitability for real‑time, high‑volume analytical workloads.