Unnecessary cut spells are glued together, overlapping spells are "polished", etc.
join.spells(dat, strict = FALSE, eps = 1e-08)A data frame with names enter, exit, event, id.
If TRUE, nothing is changed if errors in spells (non-positive length, overlapping intervals, etc.) are detected. Otherwise (the default), bad spells are removed, with "earlier life" having higher priority.
Tolerance for equality of two event times. Should be kept small.
A data frame with the same variables as the input, but individual spells are joined, if possible (identical covariate values, and adjacent time intervals).
In case of overlapping intervals (i.e., a data error), the appropriate id's
are returned if strict is TRUE.
Therneau, T.M. and Grambsch, P.M. (2000). Modeling Survival Data: Extending the Cox model. Springer.