tex.catwitherror does not work well with vectors

[kostrzewa]

There should probably be some internal loops, such that vector-valued quantities can be dealt with. At present, this fails because the conditionals don't play nice with vectors of x and dx.

[urbach]

There should probably be some internal loops, such that vector-valued quantities can be dealt with. At present, this fails because the conditionals don't play nice with vectors of `x` and `dx`.

well, it can be used with apply...!

[kostrzewa]

well, it can be used with apply...!

That's true, but it cannot be used with dplyr::mutate like so:

data <- data.frame(val = [...], err = [...])
data <- dplyr::mutate(data, tex = tex.catwitherror(x = val, dx = err))

which is a really handy function for adding new columns to data frames based on existing columns.

[martin-ueding]

I always just use mapply:

mutate(tex = mapply(tex.catwitherror, val, err))

[martin-ueding]

It would be trivial to have a the current non-vectorized version as tex.catwitherror.scalar and then have a new tex.catwitherror which just calles the implementation with mapply. The problem is that currently it is allowed to pass value and error as a two-element vector as the first argument. This would become inconsistent then.

[urbach]

It would be trivial to have a the current non-vectorized version as `tex.catwitherror.scalar` and then have a new `tex.catwitherror` which just calles the implementation with `mapply`. The problem is that currently it is allowed to pass value and error as a two-element vector as the first argument. This would become inconsistent then.

Well, rather the new one would get a different name, wouldn't it?