I currently have a python cell that reads a remote parquet file in Digital Ocean.
with fs.open(file_path, "rb") as f:
table = pq.read_table(f)
# Write the table to a buffer as a Parquet file
sink = pa.BufferOutputStream()
pq.write_table(table, sink)
parquet_data = sink.getvalue().to_pybytes()
# Write the raw bytes to stdout
sys.stdout.buffer.write(parquet_data)
And then I load it in duckdb
const db = DuckDBClient.of({ siembras: siembrasDO });
When running Observable Desktop, if I delete the file in the .observable/cache, the data is not fetched again and I get an error.
However, if I run Observable Notebook Kit (CLI) and delete those files from the same cache directory, they are retrieved again from the remote server and repopulated into the cache (as expected).
Should both environments attempt to re-fetch required data from the remote source, and repopulate .observable/cache automatically?
I currently have a python cell that reads a remote parquet file in Digital Ocean.
And then I load it in duckdb
When running Observable Desktop, if I delete the file in the
.observable/cache, the data is not fetched again and I get an error.However, if I run Observable Notebook Kit (CLI) and delete those files from the same cache directory, they are retrieved again from the remote server and repopulated into the cache (as expected).
Should both environments attempt to re-fetch required data from the remote source, and repopulate .observable/cache automatically?