А[
Size: a a a
А[
A
А[
A
R
R
А[
S
A
АК
data.table
на подходе новая сущность:New symbol available in `j`, `.NGRP`, which counts the number of groups in the current grouping, part of [#1206](https://github.com/Rdatatable/data.table/issues/1206). This is a complement to symbols already available in `j`: `.GRP` indexes the current group while `.I` indexes rows globally; `.N` gives the number of rows in the current group while `.NGRP` gives the number of groups globally.
АК
.SDcols=is.numeric now works; i.e., SDcols= accepts a function which is used to select the columns of .SD, #3950. Any function (even ad hoc) that returns scalar TRUE/FALSE for each column will do; e.g., .SDcols=!is.character will return non-character columns (a la Negate()). Note that patterns= can still be used for filtering based on the column names.
А
data.table
на подходе новая сущность:New symbol available in `j`, `.NGRP`, which counts the number of groups in the current grouping, part of [#1206](https://github.com/Rdatatable/data.table/issues/1206). This is a complement to symbols already available in `j`: `.GRP` indexes the current group while `.I` indexes rows globally; `.N` gives the number of rows in the current group while `.NGRP` gives the number of groups globally.
А
AP
АК
j
, когда у тебя уже куча всего напихана в by
.АК
uniqueN
- всё таки станет не нужен, т.к. это дополнительная операция со сканированием векторов, тогда как в .NGRP
всё будет считаться сразу и просто доступна.A
.SDcols=is.numeric now works; i.e., SDcols= accepts a function which is used to select the columns of .SD, #3950. Any function (even ad hoc) that returns scalar TRUE/FALSE for each column will do; e.g., .SDcols=!is.character will return non-character columns (a la Negate()). Note that patterns= can still be used for filtering based on the column names.
NH
NH