mbutrovich commented on code in PR #17162: URL: https://github.com/apache/iceberg/pull/17162#discussion_r3571340766
########## site/docs/blog/posts/2026-07-10-accelerating-iceberg-rust-development-with-datafusion-comet.md: ########## @@ -0,0 +1,213 @@ +--- +date: 2026-07-10 +title: Accelerating Apache Spark Queries (and Iceberg Rust Development) with Apache DataFusion Comet +slug: accelerating-iceberg-rust-development-with-datafusion-comet # this is the blog url +authors: + - mbutrovich +categories: + - blog +--- + +<!-- + - Licensed to the Apache Software Foundation (ASF) under one or more + - contributor license agreements. See the NOTICE file distributed with + - this work for additional information regarding copyright ownership. + - The ASF licenses this file to You under the Apache License, Version 2.0 + - (the "License"); you may not use this file except in compliance with + - the License. You may obtain a copy of the License at + - + - http://www.apache.org/licenses/LICENSE-2.0 + - + - Unless required by applicable law or agreed to in writing, software + - distributed under the License is distributed on an "AS IS" BASIS, + - WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. + - See the License for the specific language governing permissions and + - limitations under the License. + --> + +<!-- more --> + +Apache Iceberg provides a universal table format that serves as a foundation for modern data +lakehouse +platforms. With Iceberg, users store their tables with the benefit of being able to access +and modify their data from a number of different query engines. +[Apache Spark](https://spark.apache.org) is the engine most closely associated with Iceberg. The +[Iceberg Java repository](https://github.com/apache/iceberg), the *de facto* reference +implementation of the Iceberg specification, ships Spark as its most mature integration. It is also the +engine most teams rely on for table maintenance like compaction and snapshot expiration. +In addition to Java, the Iceberg community maintains a number +of other Iceberg implementations like [C++](https://github.com/apache/iceberg-cpp), +[Go](https://github.com/apache/iceberg-go), and [Rust](https://github.com/apache/iceberg-rust). +These other implementations benefit not only from the Iceberg specification, but also the lessons +learned and design decisions of the Java project's community. The Java repository's extensive +test suites, for instance, include nearly 10,000 correctness tests driven by Spark (as of Iceberg +1.11 with Spark 4.1). Each implementation maintains its own test suite and can look to Iceberg Java +as a reference for both correct behavior and test coverage. None of them, however, can run Java's +tests directly against their own code. + +While Spark remains a powerful and robust engine, a number of projects exist to accelerate its +JVM-backed execution. One such solution is +[Apache DataFusion Comet](https://datafusion.apache.org/comet/), which Apple donated in 2024 +as a subproject of the [Apache DataFusion](https://datafusion.apache.org) query engine. Comet's +native execution engine runs CPU-bound jobs faster and IO-bound jobs with +fewer resources, giving users control over how they want to optimize their Spark jobs. As we will +see, Comet does more than speed up queries: the same design that makes it fast also makes it a tool +for accelerating Iceberg Rust's development. + +## Accelerating Spark Queries with Comet + +Comet builds upon several related Apache projects including DataFusion (for its efficient operator +implementations like joins and aggregations), [Arrow-rs](https://github.com/apache/arrow-rs) +(for its standardized in-memory format and robust Parquet reader), and, somewhat surprisingly, Review Comment: I guess that we're loading two difference Iceberg implementations' libraries in the same process (_i.e._, one is used for planning, and the other is used for execution). That's not obvious though, so I can rephrase. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
