Due to bug https://github.com/geopandas/geopandas/issues/2495 in geopandas, we are using a non-public API until the bug is solved.
We need to set version 2.6 of arrow to be able to export datetimes with nano-second precision, which is required for lightning.
# Convert using private API until the following bug is solved:
# https://github.com/geopandas/geopandas/issues/2495
# Replace later with somethine like:
# dataframe.to_parquet(file, version = "2.6")
import geopandas as gp
table = gp.io.arrow._geopandas_to_arrow(dataframe)
import pyarrow.parquet as pq
pq.write_table(table, file, version="2.6")
Due to bug https://github.com/geopandas/geopandas/issues/2495 in geopandas, we are using a non-public API until the bug is solved. We need to set version 2.6 of arrow to be able to export datetimes with nano-second precision, which is required for lightning.
See: https://github.com/stactools-packages/goes-glm/blob/main/src/stactools/goes_glm/parquet.py#L124
with