Examples

Read and write six vector formats with downloadable Apache Hop pipelines.

Vector Reader → Vector Writer

Start with three points, inspect their fields, and export them to a format your next tool can read. This series covers all six vector output formats in the Vector Raster plugin, including the differences that matter when exchanging data.

Tested with Apache Hop 2.19.0, distribution 0.2.1-SNAPSHOT.build.35761791293.1, on macOS ARM64 with Java 21. The examples use the local pipeline engine.

Download all pipelines and sample data

Set up the example project once

Install GeoHop, then download and extract the whole example ZIP into a writable folder. Keep its data, metadata, output and pipeline files together.
In Hop GUI, use the Projects controls to create a project named vector-formats. Set its Home folder to the extracted vector-formats folder and its configuration file to the included project-config.json. Use the existing configuration when prompted.
Select that project. Open one of the .hpl files from its home folder.
Use the included local pipeline run configuration. No environment or external database is required.

All file settings use ${PROJECT_HOME} plus a relative path. The plugins expect local filesystem paths; do not substitute a file: URI. The output directory is included and must remain present.

Each main pipeline reads data/sites.gpkg, layer sites, and writes one independent output. Create examples deliberately refuse to replace an existing output. To repeat a tutorial, choose a new output name or remove only that tutorial’s generated dataset. Keep data/sites.gpkg unchanged.

Meet the sample data

These are three invented XY points in EPSG:2056 — CH1903+ / LV95. They are tutorial fixtures, not surveyed locations.

site_id	name	height_m	Easting	Northing
1	Aare	430.50	2600000	1200000
2	Brücke	432.25	2600100	1200050
3	Park	428.75	2600200	1200100

height_m is an ordinary numeric attribute, not a Z coordinate. The expected extent is (2600000, 1200000)–(2600200, 1200100). The geometry output field is named geometry. The umlaut in Brücke makes text-encoding problems visible.

Choose a format

Tutorial	Read	Write	What you will learn
Shapefile	Yes	Yes	Sidecar files, DBF names, encoding and a read-back check.
GeoPackage	Yes	Yes	Create a file, add a layer and append features.
File Geodatabase	Yes	Yes	Dataset selection, generated IDs and coordinate precision.
FlatGeobuf	No	Yes	Indexed output, feature order and empty geometry rules.
Native spatial Parquet	No	Yes	Typed columns and native GEOMETRY/GEOGRAPHY support.
ArcInfo GENERATE	No	Yes	Point records, coordinate formatting and attribute limits.

“Read” refers to Vector Reader, not to every other tool that might support the format. FlatGeobuf, Parquet and GENERATE are output-only in this plugin.

Shared settings and checks

In Vector Reader, select the file, load/select the layer, inspect its schema and set Geometry output field to geometry. In Vector Writer, select that geometry field and the desired output format. Keep one writer copy per output dataset.

Assigning a CRS does not reproject coordinates

The CRS setting describes coordinates already in the input. Assigning EPSG:4326 to these EPSG:2056 coordinates would not turn them into longitude and latitude. Reproject coordinates in a suitable upstream operation when the target CRS differs.

Check feature counts, attributes, geometry type, coordinates and CRS after conversion. A successful pipeline alone does not prove that every source characteristic survived: formats differ in their handling of Z/M dimensions, curves, empty geometries and field types.

The tutorials show real Hop dialogs from the documented distribution. General transform behaviour is covered in the Vector Raster handbook (German); additional runnable pipelines live in the repository’s examples folder.