Skip to content

write_feather

Transforms the input event stream to Feather byte stream.

write_feather [compression_level=int, compression_type=str, min_space_savings=double]

Transforms the input event stream to Feather (a thin wrapper around Apache Arrow’s IPC wire format) byte stream.

An optional compression level for the corresponding compression type. This option is ignored if no compression type is specified.

Defaults to the compression type’s default compression level.

Supported options are zstd for Zstandard compression and lz4 for LZ4 Frame compression.

:::tip Why would I use this over the compress operator? The Feather format offers more efficient compression compared to the compress operator. This is because it compresses the data column-by-column, leaving metadata that needs to be accessed frequently uncompressed. :::

Minimum space savings percentage required for compression to be applied. This option is ignored if no compression is specified. The provided value must be between 0 and 1 inclusive.

Space savings are calculated as 1.0 - compressed_size / uncompressed_size. For example, with a minimum space savings rate of 0.1, a 100-byte body buffer will not be compressed if its expected compressed size exceeds 90 bytes.

Defaults to 0, i.e., always applying compression.

load_file "input.json"
read_json
write_feather
save_file "output.feather"

to_hive, write_bitz, write_parquet