How to replace column in a dynamically-named data frame in R

Question

I'm working on some stuff for my fantasy football league. I populate a series of data frames, each named for the position (e.g., QB for quarterbacks, RB for running backs, etc.)

I've been able to use the apply() function to do a lot with rbinding the separate chunks of a position into a single dataframe, but what I'd really like to do is to strip out some text from something for each position.

So I have a dataframe named QB, and one of the columns is called Player, that has some superfluous text I'd like to remove.

So, if I were to do this, it works:

thePosNick <- "QB"
thePlayer <- paste(thePosNick, "$Player", sep="")
g <- str_replace(eval(parse(text=thePlayer)), "remove this text", "")
g <- str_replace(g, "more meaningless text", "")
g <- str_trim(g)
#expected behavior:
QB$Player <- g

That works fine, however, I'd love to be able to do this dynamically, and doing the following doesn't seem to work:

thePosNick <- "QB"
thePlayer <- paste(thePosNick, "$Player", sep="")
g <- str_replace(eval(parse(text=thePlayer)), "remove this text", "")
g <- str_replace(g, "more meaningless text", "")
g <- str_trim(g)
#but i'd love to be able to do something like this:
assign(thePlayer, value=g)

Can anyone tell me what I'm doing wrong? Any suggestions are greatly appreciated, thanks!

Answer 1

First, I'd make a function to do your fixing.

fixit <- function(g) {
  g <- str_replace(g, "remove this text", "")
  g <- str_replace(g, "more meaningless text", "")
  g <- str_trim(g)
  g
}

Then get the variable you want, fix it, and assign back.

thePosNick <- "QB"
x <- get(thePosNick)
x$Player <- fixit(x$Player)
assign(thePosNick, x)

Alternatively, a more R idiomatic way would be to use a list to store all your positions instead of just storing them in the global environment. Then you could use lapply to fix them all at once.

positions <- list(QB=QB, RB=RB, TE=TE)
positions <- lapply(positions, function(x) { x$Player <- fixit(x$Player) } )

Answer 2

Store your stuff like this?

biglist = list(QB=list(Player='yo remove this text here'))

# $QB
# $QB$Player
# [1] "yo remove this text here"

Then

g <- biglist[[thePosNick]][['Player']]
biglist[[thePosNick]][['Player']] <- str_replace(g, "remove this text", "")

# $QB
# $QB$Player
# [1] "yo  here"

That is, you can put names of things inside [[]]. You're only encountering trouble assigning to stuff in the global environment (as opposed to nested in a list), I reckon.

Comment. As @Roland said, you'd be better of yet storing this in a data.frame or data.table, perhaps with one row per player: DF <- data.frame(Player='You',Pos='QB',Team='Youse',...)

Answer 3

You must use

eval(parse(text=paste(thePlayer,"<-",deparse(g))))

instead of assign(thePlayer, value=g).

Why? Because QB$Player <- g is not a direct assignment, but a call to the function '$<-', which places the value in g in a special position of the object QB. There is no object QB$Player in your environment, that is why assign() fails.

Opinion:

This is all pretty bad because the better solution is to use a list of dataframes like this:

L <- list(QB=QB, RB=RB, ...)

g <- str_replace(L[["QB"]]$Player, "remove this text", "")
g <- str_replace(g, "more meaningless text", "")
g <- str_trim(g)
L[["QB"]]$Player <- g

Answer 4

One quick and dirty solution would be

eval(parse(text=paste(thePlayer,"<-g",sep="")))

I can't help but think, though, manipulating these objects by their names as strings might not be the cleanest way to go about this whole problem.