Conversation
Codecov Report
@@ Coverage Diff @@
## master #3081 +/- ##
==========================================
+ Coverage 90.86% 90.88% +0.02%
==========================================
Files 61 61
Lines 11808 11804 -4
==========================================
- Hits 10729 10728 -1
+ Misses 1079 1076 -3
Continue to review full report at Codecov.
|
|
|
||
| S3method(anyDuplicated, data.table) | ||
|
|
||
| export(split.data.table) |
There was a problem hiding this comment.
is this really required? this was exported in 1.11.6 #2920 as per user request in a comment somewhere. As it is very recent it won't break much code, but NEWS entry would be useful.
There was a problem hiding this comment.
A comment next to it explaining why and when exported would have been good because it's not proper to do so in addition to the S3method(split, data.table) which is still there. Should we consider exporting all methods for base generics as well as registering them?
Closes #3080
Completes final item in #2740
Added comments to the top of
fmelt.cas to why we don't importreshape2(R dependency inheritance)meltis generic inreshape2so I'm not sure why we were explicitly calling itdcastdispatches non-data.table to non-genericreshape2::dcastthe same as before, but with a more helpful message ifreshape2is not installed.Tidied NAMESPACE. There were a few methods explicitly exported which didn't need to be since S3method already exports the method. This leaves just one use of the new delayed S3method registration, for
xts.The
as.Datemethod has another twist in thatzoo::as.Datemasksbase::as.Dateand they are both generic.