Future SDTL Developments

  1. Extend SDTL to include data created by statistical analysis commands, such as predicted values computed from a regression model.

  2. Add support for variable types, such as numeric versus text. The current model does not require variable type information. But there are a number of situations where knowing the variable type would make a difference. For example, in Python the equivalent of an SDTL Collapse command can be run on all variables in the dataframe. If the aggregate to be created is a sum or average, only numeric variables will be included in the new dataframe. We currently cannot identify which variables are numeric. Support for variable types will require a number of changes in SDTL, including adding variable type to the outcomes of the Function Library.