TODO.md

In decisionpatterns/ordering: Test, Check, Verify, Investigate the Monotonic Properties of Vectors

TODO

• [ ] Change package name to sorted (sordid?) or order

• [ ] Add step_size() for reporting step size for incremental vectors.

• See base::is.unsorted. There can be many uses of this but can probably for example 3:0 is unsorted

• [ ] Should is_incremental when applied to non-numerics:

• error (current)
• return FALSE

• return FALSE with warning

• Covnert to C++ for speed

• Performance not ....

' Instead of using diff, \code{is_increasing} and \code{is_decreasing} uses

' \code{cummax} and \code{cummin} respectively for efficiency. See the

' stackoverflow reference for timings and an explanation.

CLOSED

• Add
• [ ] is_not_increasing
• [ ] is_not_strictly_increasing

• etc. This is just ! is_strictly_increasing(x)

• [x] [LOW:switch to base::is.unsorted] is(strictly)[increasing|decreasing) and is_sorted should:

• be implemented in cpp
• be efficient by comparing sequence of values

• [x] support all types

• [ ] Provide options for 'monotonicity' to report "increasing" , "strictly_increasing", etc. maybe as a factor.

  1:10 %>% monotonicty()

• [x] Should most of the functionality of the package be moved to use [base::is.unsorted()] as this seems to be a fast/standard method?

• [x] Should we support not_ functions? NO. This is confusing using %>% not() is cleaner

• [x] Should the argument be na.rm rather than na.action to match [base::is.unsorted(). Are there other types of na.actions that can be applied.

• What to with ----- vectors (e.g. constants) vs /__/ (non-monotonic)

• [x] Add is_constant

decisionpatterns/ordering documentation built on May 5, 2019, 7:09 a.m.