Create an exp_data or a val_data

Create a experience or valuation data from a list of columns, a data.frame or tibble (or similar).

The columns must be atomic, the same length and uniquely named.

There must be a birth date column.

A count column is optional. If present it must be numeric, non-negative and finite. Rows with a nil count are dropped.

Experience data must also contain (and valuation data must not contain)

• E2R_start and E2R_end date columns, and

• an E2R_died logical column.

These special columns cannot contain NA and, if they are dates, they must be datey.

The experience data period and individual E2Rs must be proper:

• at the dataset level, exp_start cannot be after exp_end, and

• at the record level, E2R_start cannot be after E2R_end and if they are the same then E2R_died must be false.

The experience data is trimmed to the experience period, i.e. it is guaranteed that E2R_start is not before exp_start and E2R_end is not after exp_end. Records with empty E2Rs after trimming are dropped.

Usage

exp_data(
  columns,
  exp_start,
  exp_end,
  on_unreadable = c("error", "warn", "ignore")
)

val_data(columns, as_at, on_unreadable = c("error", "warn", "ignore"))

Arguments

columns

The list of columns (which includes things like data.frame and tibble) to convert to an exp_data or val_data.

exp_start

The start of the overall experience period (inclusive). Must be before exp_end.

exp_end

The end of the overall experience period (exclusive). Must be after exp_start.

on_unreadable

What to do about columns whose type logmu cannot read (anything other than double, integer, logical, character, factor, datey or durationy): "error" (the default) rejects them, "warn" keeps them with a warning, and "ignore" keeps them silently. The special columns (birth, the E2R_* columns, count) are always checked.

as_at

The 'as at' date of the valuation data.

Value

An exp_data or val_data.