using .by means the output will have one row by set of grouping columns
grouping columns can be selected using the same principles as what has been described in https://github.com/moodymudskipper/tb/issues/5, we can even define new columns on the spot using s()
With data.table if non NULLby:= in j means "mutate by" (by ref), and .() in j means "summarize by" (by copy).
In tb.by always mean summarize in the wide sense (i.e. spreading is a summarizing operation too). To "mutate by" we'll use a 2 sided formula (see https://github.com/moodymudskipper/tb/issues/9).
If .by is not missing but of length 0 (NULL or empty vector/list), The result summarized without groups, ``
s()With data.table if non
NULLby:=injmeans "mutate by" (by ref), and.()injmeans "summarize by" (by copy).In
tb.byalways meansummarizein the wide sense (i.e. spreading is a summarizing operation too). To "mutate by" we'll use a 2 sided formula (see https://github.com/moodymudskipper/tb/issues/9).If
.byis not missing but of length 0 (NULL or empty vector/list), The result summarized without groups, ``