Combine settings

Configuration for combining ORD results (combine_settings_schema.json), used when grouping and aggregating result tables across analyses.

Field	Type	Required	Constraints	Description
occ_dtype	array[array]			dtype for records in the occurrence.bin file
group_fill_perspectives	boolean		default false	Enable filling financial perspectives to enable more groupings of output sets.
group_event_set_fields	array[string]		each item one of: "event_set_id", "event_set_description", "event_occurrence_id", "event_occurrence_description", "event_occurrence_max_periods", "model_supplier_id", "model_name_id", "model_description", "model_version"	List of fields
group_number_of_periods	integer	Yes		Number of periods for grouping
group_period_seed	integer		default 2479	Seed for the grouping random number generator.
group_correlation	float			Correlation to apply to the group.
group_mean	boolean		default false	Output the grouped mean results.
group_mean_type	enum[integer]		one of: 1, 2	Output the grouped mean results.
group_secondary_uncertainty	boolean		default false	Output the grouped results with secondary uncertainty (type 2)
group_parametric_distribution	enum[string]		one of: "beta", "gamma"	The parameteric distribution type to estimate values for from MELT.
group_format_priority	array[string]		each item one of: "M", "Q", "S"	Control which format for ELTs is used for loss severity sampling when there are multiple formats available per output set. M = mean, Q = quantile, S = sample. ['M'] means only MELTs will be used for loss sampling. ['Q', 'M'] means that QELT will ber prioritised, falling back to MELT if QELT format is missing.
group_alt	boolean		default false	Output grouped ALT report
group_ept	boolean		default false	Output grouped EPT report
group_ept_oep	boolean		default true	Output grouped EPT with OEP if group_ept
group_ept_aep	boolean		default true	Output grouped EPT with AEP if group_ept
group_plt	boolean		default false	Output grouped PLT report
output_type	string		default "csv"	Output type for all outputs. Options are ‘csv’ or ‘parquet’