apache / hudi

Apache Hudi Apache Hudi is an open data lakehouse platform, built on a high-performance open table format to ingest, index, store, serve, transform and manage your data across multiple cloud data environments. Features Hudi stores all data and metadata on cloud storage in open formats, providing the following features across different aspects. Ingestion • Built-in ingestion tools for Apache Spark/Apache Flink users. • Supports half-dozen file formats, database change logs and streaming data systems. • Connect sink for Apache Kafka, to bring external data sources. Storage • Optimized storage format, supporting row & columnar data. • Timeline metadata to track history of changes • Automatically manages file sizes, layout using statistics • Savepoints for data versioning and recovery • Schema tracking and evolution. Indexing • Scalable indexing subsystem to speed up snapshot queries, maintained automatically by writes. • Tracks file listings, column-level and partition-level statistics to help plan queries efficiently. • Record-level indexing mechanisms built on row-oriented file formats and bloom filters. • Logical partitioning on tables, using expression indexes to decouple from physical partitioning on storage. Writing • Atomically commit data with rollback/restore support. • Fast upsert/delete support leveraging record-level indexes. • Snapshot isolation between writer & queries. • Optimistic concurrency control to implement relational data model, with Read-Modify-Write style consistent writes. • Non-blocking concurrency control, to implement streaming data model, with support for out-of-order, late data handling. Queries Hudi supports different types of queries, on top of a single table. • **Snapshot Query** - Provides a view of the table, as of the latest committed state, accelerated with indexes as applicable. • **Incremental Query** - Provides latest value of records inserted/updated, since a given point in time of the table. Can be used to "diff" table states between two points in time. • **Change-Data-Capture Query** - Provides a change stream with records inserted or updated or deleted since a point in time or between two points in time. Provides both before and after images for each change record. • **Time-Travel Query** - Provides a view of the table, as of a given point in time. • **Read Optimized Query** - Provides excellent snapshot query performance via purely columnar storage (e.g. Parquet), when used with a compaction policy to provide a transaction boundary. Table Management • Automatic, hands-free table services runtime integrated into Spark/Flink writers or operated independently. • Configurable scheduling strategies with built-in failure handling, for all table services. • Cleaning older versions and time-to-live management to expire older data, reclaim storage space. • Clustering and space-filling curve algorithms to optimize data layout with pluggable scheduling strategies. • Asynchronous compaction of row oriented data into columnar formats, for efficient streaming writers. • Consistent index building in face of ongoing queries or writers. • Catalog sync with Apache Hive Metastore, AWS Glue, Google BigQuery, Apache XTable and more. Learn more about Hudi at https://hudi.apache.org Building Apache Hudi from source Prerequisites for building Apache Hudi: • Unix-like system (like Linux, Mac OS X) • Java 11 or 17 • Git • Maven (>=3.6.0) To build for integration tests that include , use . To build the Javadoc for all Java and Scala classes (project should be already compiled): Build with different Spark versions The default Spark 3.x version, corresponding to profile is 3.5.3. The default Scala version is 2.12. Scala 2.13 is supported for Spark 3.5 and above. Refer to the table below for building with different Spark and Scala versions. | Maven build options | Expected Spark bundle jar name | Notes | |:--------------------------|:---------------------------------------------|:-------------------------------------------------| | (empty) | hudi-spark3.5-bundle_2.12 | For Spark 3.5.x and Scala 2.12 (default options) | | | hudi-spark3.3-bundle_2.12 | For Spark 3.3.2+ and Scala 2.12 | | | hudi-spark3.4-bundle_2.12 | For Spark 3.4.x and Scala 2.12 | | | hudi-spark3.5-bundle_2.12 | For Spark 3.5.x and Scala 2.12 (same as default) | | | hudi-spark3.5-bundle_2.13 | For Spark 3.5.x and Scala 2.13 | | | hudi-spark4.0-bundle_2.13 | For Spark 4.0 and Scala 2.13 (Needs java 17) | | | hudi-spark3-bundle_2.12 (legacy bundle name) | For Spark 3.5.x and Scala 2.12 | Please note that only Spark-related bundles, i.e., , , , can be built using profile. Hudi Flink bundle cannot be built using profile. To build these bundles on Scala 2.13, use the following command: For example, What about "spark-avro" module? Starting from versions 0.11, Hudi no longer requires to be specified using Build with different Flink versions The default Flink version supported is 1.20. The default Flink 1.20.x version, corresponding to profile is 1.20.1. Flink is Scala-free since 1.15.x, there is no need to specify the Scala version for Flink 1.15.x and above versions. Refer to the table below for building with different Flink and Scala versions. | Maven build options | Expected Flink bundle jar name | Notes | |:--------------------|:-------------------------------|:---------------------------------| | (empty) | hudi-flink1.20-bundle | For Flink 1.20 (default options) | | | hudi-flink2.1-bundle | For Flink 2.1 | | | hudi-flink2.0-bundle | For Flink 2.0 | | | hudi-flink1.20-bundle | For Flink 1.20 (same as default) | | | hudi-flink1.19-bundle | For Flink 1.19 | | | hudi-flink1.18-bundle | For Flink 1.18 | | | hudi-flink1.17-bundle | For Flink 1.17 | For example, Running Tests Unit tests can be run with maven profile . Functional tests, which are tagged with , can be run with maven profile . Integration tests can be run with maven profile . To run tests with spark event logging enabled…