modifying: Modify part of a value.
In crowding/vadr: Macros, metaprogramming, and pleasant programming idioms.

Description Usage Arguments Details Value Author(s) See Also Examples

The macro put provides syntax for modifying part of an object in a functional context (i.e. creating a modified value without necessarily binding the result to a name.) Unlike <-, the value of the expression is the modified object, not the value that was injected. This is particularly useful in combination with 'chain.'

alter takes the selected subset of it, then filters it through additional functions in the manner of chain, then replaces the subset with the result, returning the modified object.

inject takes the entire object, filters it through a chain, then places the result in the specified subset.

put(it, subset, value)

alter(it, subset, ...)

inject(it, subset, ...)

`...`	A `chain` of code transformations.
`it`	A value.
`subset`	A subassignment target expression; this is interpreted literally if the symbol `it` is used, otherwise `it` is injected as far down the leftmost arguments of the expression as possible. (Thus `names` is interpreted as `names(it)`, and `names[1]` as `names(it)[1]`.)
`value`	The value to assign

Normal subassignment in R is effectively a macro, one which turns a statement like

names(x)[1] <- "head"

into something like

x <- `names<-`(x, `[<-`(names(x), "head", 1))

However even this explanation is misleading, because the value returned from a subassignment is the value applied, not the value assigned. Consider if you wanted to call a function with a modification of an existing value:

do_something(names(x)[1] <- "head")

Aside from changing the value of x this actually doesn't pass the value to do_something, but rather performs the equivalent of do_something("head").

In this situation, using put, one can write:

do_something(put(x, names[1], "head"))

codeput and friends are particularly useful in conjunction with chain.

x %<~% alter(names[5], toupper) is equivalent to:

names(x)[5] <- toupper(names(x)[5])

x <- inject(1:10, names, letters[.], toupper) is equivalent to:

x <- 1:10; names(x) <- toupper(letters[x])

The modified value.

Peter Meilstrup

chain

put(1:10, names, letters[1:10])
x <- 1:10
put(x, it[1], 4)
put(x, names[4], 'a')
x #x is unchanged
x <- alter(structure(1:10, names=letters[1:10]), names)
y <- alter(x, names[5], toupper, str_dup(3))
x <- inject(1:10, names[1:5], letters[.], rev)