The tbl_es class is a subclass of dtplyr::lazy_dt() created in order to
store attributes for experience studies. The colloquial term 'expstudy'
refers to a lazy_dt() that has the tbl_es subclass.
Properties of tbl_es
Attributes that store variable names used for key metric calculations:
actuals: the observed variable (or variables) within the experience study (e.g., actual lapse counts for a lapse study).expecteds: the expected variable (or variables) within the experience study (e.g., expected number of deaths for a mortality study).exposures: the exposure variable (or variables) within the experience study. This will be the time the records spans in years (or within (0, 1) for records covering less than a year).variances: optional, the variance variable (or variables) within the experience study. This may not be available for all studies, thus can be omitted. Variance is used for the credibility metric.
An additional (optional)
keysattribute that can uniquely describe every record within the experience study. This typically is policy number and a measure of time, such as annual or monthly duration. Can be omitted, but providing enables quicker binary searches within thedtplyr::lazy_dt(). For more information, refer to the data.table vignette:vignette('datatable-keys-fast-subset', package = 'data.table').
Behavior of tbl_es
A tbl_es behaves the same as a dtplyr::lazy_dt() only differing by
retaining the attributes listed above. This saves time in routine, repetitive
coding, leading to increased time for analysis.