The batch companion of load_scenario(): resolves scenario NAMES through
the project registry (all registered scenarios by default) and loads each
as a thin on-disk shell. The result — a named list, or a filled
environment — feeds getData() directly:
getData(load_scenarios(), name = "vTechOut", merge = TRUE).
A registered scenario whose folder is gone is skipped with a warning
(run refresh_registry() to drop the stale row).
Usage
load_scenario(
path,
name = NULL,
env = .scen,
overwrite = FALSE,
ignore_errors = FALSE,
verbose = TRUE
)
load_scenarios(names = NULL, run = NULL, env = NULL, verbose = TRUE)Arguments
- path
character. Path to saved with function
save_scenarioscenario directory.- name
character. Name to assign to the loaded scenario object. By default, the name is taken from the loaded scenario object.
- env
environment to assign the scenarios into (e.g.
.scen), orNULL(default) to return them as a named list.- overwrite
logical. Overwrite existing scenario object in the environment.
- ignore_errors
logical. Ignore load errors and continue execution. This option is useful when some data is missing or corrupted.
- verbose
logical.
- names
character, scenario names to load;
NULL(default) = every scenario in the registry.- run
character, optional run id to activate in each loaded scenario via
read_solution()(scenarios without that run keep their default, with a warning).
Value
TRUE if scenario is loaded, FALSE if not.
a named list of scenarios, or (with env=) the loaded names,
invisibly.
Slots stay on disk
A loaded scenario keeps its data-frame slots in the Arrow store rather than
in memory, so reading one directly — obj@demand, obj@capacity — returns
an empty frame rather than the data, and isInMemory() is FALSE.
getData() goes through the lazy store and is unaffected, so it is the
route to use; findData() and get_variable() likewise. This bites report
and plotting code written against an in-memory model, where the symptom is
an empty table rather than an error.
