RatZarr 1.0.2 has been released

This release supports the use of multi-valued columns. In effect, this allows for a single column to be a two dimensional array instead of the usual one dimension. The second dimension allows for multiple values to be stored for each row of the RAT. Previously, the only way to do this was to store them as separate columns, and rely on the column names to indicate that they are closely related columns. With this new feature, they are stored as a single RAT column, and read in as a single 2-d array

It is expected that most columns will remain 1-d. However, some suggested use cases for 2-d columns include:

  • Lidar waveform data for each image segment
  • Multi-year FPC (Foliage Projective Cover) values per-segment

It is strongly recommended that when using this multi-valued column feature, a column attribute also be stored (see getColumnAttributeMapping) to document the meaning of the sub-columns. For example, if storing FPC values for multiple years as a single column named FPC, one might also store a column attribute with the list of years, e.g.

attrs = rz.getColumnAttributeMapping('FPC')
attrs['years'] = [1985, 1990, 1995, 2000, 2005]