Add filter.mids and apply in mice.mids(., newdata = df)

[prockenschaub]

Scope

This pull request adds a filter.mids function as discussed in #32 and #266 .

Progress

• Implement filter.mids

• Create tests for ignore in filter.mids

• Create documentation for filter.mids

• Apply filter to mice.mids(., newdata = df)

[prockenschaub]

The functionality implemented here is just the basice functionality, allowing to filter the object based on a logical include vector. It should be relatively easy to also allow for arbitrary selections based on the imputations by accepting a closure for include, which is evaluated within filter.mids and passed the mids object .

[stefvanbuuren]

Thanks. This is great.

Commit 54efb0c convert filter.mids() into an S3-method for dplyr::filter(), so we can now also use syntax like

imp <- mice(nhanes, m = 2, maxit = 1, print = FALSE)
imp_f2 <- filter(imp, age >= 2 & hyp == 1)
nrow(complete(imp_f2))  # should be 5

Now merged into add_ignore.