Ethan Smith
Ethan Smith
i've taken a stab at the unit test and changes to `first.r` https://github.com/joshuaulrich/xts/compare/master...ethanbsmith:fix_350 could you take a look and give any feedback before i start on `last.r`
edge case test for all inputs having zero rows ```r x
some color on my perspective: 1. it conforms to relational algebra, which is how i think about R matrix merging/subsetting 2. it requires less exception handling, as the empty output...
testing on https://github.com/joshuaulrich/xts/commit/a3e9fb2b56bb1e2cfcc9067d845b064a1fe99276 ```r x
those issues all came up running through my analytics suite. will add tests if i think of anything else current patch works for all my scenarios! Thank you!
dont think the data itself should ever have values 1`, then i probably reluctantly agree
this will return different types based on the length of the output (list vs xts) would be nice to be able to have it always return a list, regardless of...
just thinking out loud: ideally, `getSymbols` would always return a list and `getSymbol` (singular) would wrap that with an `unlist()`, and auto.assign and env would get deprecated. if thats an...
also, not sure if this should be a standalone function in R, or just add a `trim` parameter to `runMean`, like how base R does it I *think* that one...
agree this should ideally just extend runMean util i get time to work on a proper C pr, here is a usable R implementation: ```r runTrimmedMean