my_aaply: Temporary local version of 'aaply'
In LAJordanger/leanRcoding: Leaner R Coding

Description Usage Arguments Details Value

The 'aaply' function did for some reason not want to cooperate for one of the cases I needed to investigate, so an ad-hoc solution had to be created in order to get a work-around. The result of this function has been sanity-checked with the result from 'aaply' for one of the cases where 'plyr::aaply' did work. Note that even though the order of the dimensions are ensured to be the same, the internal order in some of the dimensions might differ, so use restrict_array to permute them into the same order.

1	my_aaply(.arg_grid, .fun, .new_dnn = "values", .new_dn = NULL, ...)

`.arg_grid`	A data-frame, typically created by `expand.grid` used on a list containing references to the dimension names of some array containing values which we want `.fun` to work upon.
`.fun`	The function that we want to use. The first argument of `.fun` must be able to digest a row from the data-frame `.arg_grid`.
`.new_dnn`	The name to be used on the dimension of the result that originates from `.fun`.
`.new_dn`	A vector to be used upon the result that originates from `.fun`. The default value `NULL` will trigger the use of the names returned by `.fun`. The main intention of this function is to add a name when the result has length one and no names.
`...`	Arguments to be given to `.fun`.

Note that the result from functions that returns matrices/arrays probably will look like shit if used in this first coarse approximation. It might be possible to create a more advanced version of this function by using 'abind' instead of 'rbind' in the construction, but as I at the moment don't need such funcitonality, I will simply ignore it for the time being.

It might perhaps be possible to adjust this function in order to do bigger chunks when sending stuff to the available cores, which might reduce the overhead a bit. But this will require another design of the function, which I can't use time on.

For functions that return a vector of values, the returned object will be similar to the result obtained by plyr::aaply, with the difference that the new dimension will be named "values".

LAJordanger/leanRcoding documentation built on Feb. 27, 2020, 4:42 p.m.