Open J-Moravec opened 2 hours ago
All of this sounds useful and sensible.
Note that there is already support for various rbind()
strategy via the rbind2()
S4 method. See:
https://vincentarelbundock.github.io/tinytable/vignettes/tinytable.html#combination-and-exploration
We need to use rbind2
because it allows extra arguments, whereas rbind
does not (IIRC).
In addition, we might want to implement methods for:
sort_by()
All of this sounds useful and sensible.
Note that there is already support for various
rbind()
strategy via therbind2()
S4 method. See:https://vincentarelbundock.github.io/tinytable/vignettes/tinytable.html#combination-and-exploration
We need to use
rbind2
because it allows extra arguments, whereasrbind
does not (IIRC).
Sounds like all we need to do is define custom S3 method that calls rbind2 to have nicer interface. And I need to read more on S4.
I think I tried and rbind()
and cbind()
were not possible. It will have to be rbind2.tinytable()
and cbind2.tinytable()
. FWIW, I think the current interface for rbind2()
is already quite "clean" and flexible.
It should be possible to define subset.tinytable()
and sort_by.tinytable()
.
Currently, this behaviour is undocumented and in some cases, unimplemented (or it returns nonsense).
I would like to:
rbind
, just the default behaviour adds an additional headerThis would allow preparing tables as chunks and adding them together, instead of having to prepare them all at once.
User Case
I am currently preparing tables for two datasets, the code for the tables (including images) is identical, since they are summarizing the table in the same way.
But without the ability to merge them, I need to do quite a lot of complex wrangling instead to force them into
tinytable
format and prepare everything as a single big table.