Getters / setters for the codec of the files exchanged with the JuMP and
Pyomo backends: the model data written into the run's solver/input/, and
the solution written back into solver/output/. Julia's Arrow.jl reads
IPC only, so "parquet" is available on the Pyomo path and refused on the
Julia one.
These files live for the duration of one solve, which is why the compression
default is cheaper here than for storage (get_storage_compression()).
A solver preset's export_format / import_format overrides these options
for that solve, and the two directions stay independent – Arrow input with
CSV output is a legal combination. Setting a preset's export_format to
"SQLite" (Pyomo) or "RData" (Julia) selects the legacy exchange.
Usage
get_exchange_format()
set_exchange_format(format = c("feather", "parquet", "csv"))
get_exchange_compression()
set_exchange_compression(codec = c("zstd", "lz4", "uncompressed"))
get_exchange_compression_level()
set_exchange_compression_level(level = 15L)
Arguments
- format
one of "feather", "parquet", "csv".
- codec
compression codec, e.g. "zstd", "lz4", "uncompressed".
- level
integer compression level (ZSTD: 1-22; ignored for lz4).
Value
the option value (getters) or the previous value, invisibly (setters).
See also
Other options:
en_config_path(),
en_config_show(),
get_default_solver(),
get_storage_format(),
isVerbose(),
set_log_file(),
set_option(),
set_path_builder(),
set_registry_file(),
set_reports_path(),
set_scenarios_path(),
set_solver_path()
Examples
get_exchange_format()
#> [1] "feather"
