Using geoparquet with R

Overview

Using geoparquet as a file format with spatial data in R is a relatively new option. As spatial formats have evolved, there is generally a balance between read and write speed, and file size. There are differences between raster data and vector data, and primarily the focus of this presentation is on vector based data.

File Formats

Spatial formats in R for vector data can be a wide variety of types, from shapefiles, kml, or simple csv of X/Y points. In R we can access and use these data with a number of packages, from the older {sp} package (see here for more background) to the more recent and widely embraced {sf} package. For more info on the types of data that can be imported or exported via {sf}.

.parquet format

Enter another option: .parquet

This is not a new format, but it is new to use within the geospatial/R world. The parquet format is a file type that contains a table inside similar to a .csv with these differences:

• However these files are stored in binary form not as plain text
• parquet files are column-oriented (unlike csv) and each column is stored independently
• parquet embeds the schema or data types/structure within the data itself

Columnar Data

The real benefit:

• parquet files aren’t just about compression (though some savings here)
• The real benefit is speed reading/operating/writing data!

Let’s try it out! Check out these slides.