Skip to content

Modules

This page lists the artifacts published by Parquetry, with what each one is for and its status. Use it to choose the dependencies of your project.

The artifacts share one version number. Getting started gives the latest release and shows how to import the BOM.

Artifacts

Declare the Parquetry modules called by your code. Maven resolves the modules below them, such as parquetry-format under parquetry-core.

Group id Artifact id Use it to Status
io.tileverse.parquetry parquetry-core Read and write single Parquet and GeoParquet files, filter rows, and decode geometries to JTS. stable
io.tileverse.parquetry parquetry-catalog Read a directory or a glob of Parquet files as one dataset. stable
io.tileverse.parquetry parquetry-geotools Serve GeoParquet files as GeoTools features, through a GeoTools data store. stable
io.tileverse.parquetry parquetry-arrow Export a read as an Arrow IPC stream or through the Arrow C Data Interface. stable
io.tileverse.parquetry parquetry-iceberg Read Apache Iceberg tables as datasets. experimental
io.tileverse.parquetry parquetry-stac Read the GeoParquet files of a STAC collection as one dataset. experimental
io.tileverse.parquetry parquetry-avro Read and write Avro container files, the format of Iceberg manifests and manifest lists. supporting
io.tileverse.parquetry parquetry-jackson Render the rows of a read as JSON. supporting
io.tileverse.parquetry parquetry-io Read byte ranges from files and from tileverse-storage, and pool the native buffers of reads. foundation
io.tileverse.parquetry parquetry-compression Compress and decompress Parquet pages and Avro blocks. foundation
io.tileverse.parquetry parquetry-format Decode and encode the Parquet footer and page headers, and model the schema and the GeoParquet metadata. foundation
io.tileverse.parquetry.apps par Inspect, filter and copy files from the command line. experimental
  • Stable parts are ready for production use. Experimental parts work, but their API and behavior may still change. The home page explains both labels.
  • Supporting modules have no status of their own. parquetry-iceberg depends on parquetry-avro, and the par tool depends on parquetry-jackson.
  • Foundation modules sit under parquetry-core, which depends on all three. Declare parquetry-core, and Maven adds them.
  • parquetry-geotools also registers the Iceberg and STAC GeoParquet stores. These two stores read through parquetry-iceberg and parquetry-stac, and are experimental like them.
  • par is not published to Maven Central. It builds from the source repository, as The par command line tool shows.

BOMs

Two BOMs keep versions in line. Import them in the dependencyManagement section of your POM, with <type>pom</type> and <scope>import</scope>.

Group id Artifact id Pins the versions of
io.tileverse.parquetry parquetry-bom The Parquetry modules listed above, except par.
io.tileverse.parquetry parquetry-dependencies The third-party libraries used by Parquetry, such as tileverse-storage, JTS and Jackson.

Import parquetry-bom alone to pin the Parquetry modules. Import parquetry-dependencies as well to also align the versions of the libraries shared by your application and Parquetry.

Which module for which task

You want to Declare
Read or write one Parquet or GeoParquet file parquetry-core, plus a tileverse-storage module for object storage (Cloud storage)
Read many files as one dataset parquetry-catalog
Publish GeoParquet through GeoTools or GeoServer parquetry-geotools (GeoTools and GeoServer)
Hand rows to an analytics engine as Arrow parquetry-arrow (Arrow interop)
Print rows as JSON parquetry-jackson (JSON)
Read Avro container files parquetry-avro (Avro)
Read an Iceberg table or a STAC collection parquetry-iceberg or parquetry-stac (Experimental parts)

Start the JVM of your application with the flags listed in Getting started. Some modules use Java 25 preview features and fail to load without them.