R help - Mar 2008 - looping through data frames in a workspace

All,

I have a workspace containing only data frame objects. I would like to
loop though each one and clean-up text columns in them. How can I have
R loop through the list? I have tried to find an answer in R help but
the closest solution I can find is to make a static list of data
frames, as illustrated in this recent post:

---------begin post
On Tuesday 19 February 2008 (19:51:15), TLowe wrote:
> Hey Folks,
>
> Could somebody show me how to loop through a list of dataframes?  I want to
> be able to generically access their elements and do something with them.
>
> For instance, instead of this:
>
> df1<- data.frame(x=(1:5),y=(1:5));
> df2<- data.frame(x=(1:5),y=(1:5));
> df3<- data.frame(x=(1:5),y=(1:5));
> plot(df1$x,df1$y);
> plot(df2$x,df2$y);
> plot(df3$x,df3$y);
>
> I would like to do something like:
> (pseudocode)
> dfarray[1] = df1
> dfarray[2] = df2
> dfarray[3] = df3
> for (each i in dfarray) {
>   plot(i$x, i$y);
> }
It's surprisingly simple:

for(df in list(df1,df2,df3)) plot(df$x,df$y)

Best,

Martin
---------------------end post

I would like to avoid having to type-out a very long list over and
over again. I have tried every variation I could think of similar to:

for(df in list(noquote(ls()))) {

     do stuff with df

     }

I know this has to be possible. What am I missing?

Many Thanks-

On Monday 03 March 2008 (23:28:05), lucy b wrote:
> I would like to avoid having to type-out a very long list over and
> over again. I have tried every variation I could think of similar to:
>
> for(df in list(noquote(ls()))) {
>
> ? ? ?do stuff with df
>
> ? ? ?}
>
> I know this has to be possible. What am I missing?
I guess something like the following will do what you want,
but I am not completely sure ...

for(dfname in ls()){
 eval(substitute({
		<do stuff with df>
		},
      		list(
		  df=as.name(dfname)
		  )
     		)
      )
	
}

HTH,
Martin

lucy b wrote:
> 
> I have a workspace containing only data frame objects. I would like to
> loop though each one and clean-up text columns in them. How can I have
> R loop through the list? I have tried to find an answer in R help but
> the closest solution I can find is to make a static list of data
> frames, as illustrated in this recent post:
> ...
> I would like to avoid having to type-out a very long list over and
> over again. I have tried every variation I could think of similar to:
> 
> for(df in list(noquote(ls()))) {
> 
>      do stuff with df
> 
>      }
> 
If you name your data frames consistently (e.g. df1, df2, df3 etc.), then
you can access them using mget:
df1 <- data.frame(a=runif(10), b=letters[1:10])
df2 <- data.frame(c=rnorm(5))
df3 <- data.frame(11:20)
dataframes <- mget(paste("df", 1:3, sep=""),
envir=.GlobalEnv)

Alternatively, if you want every dataframe in your workspace, try:
vars <- ls()
nvars <- length(vars)
dataframes <-list()
j <- 1
for(i in 1:nvars)
  {
    if(class(get(vars[i]))=="data.frame")
      {
        dataframes[[j]] <- get(vars[i])
        names(dataframes)[j] <- vars[i]
        j <- j+1
      }
  }

Regards,
Richie.

Mathematical Sciences Unit
HSL





-- 
View this message in context:
http://www.nabble.com/looping-through-data-frames-in-a-workspace-tp15815971p15823720.html
Sent from the R help mailing list archive at Nabble.com.

Thanks for helpful replies. Unfortunately, I didn't mention that I
want to overwrite the existing data frame with the corrected one using
the same name. Here's what I have:

## get names of data frames
frames <- names(Filter(function(x) x=="data.frame",
                   sapply(objects(),function(x)
                          class(eval(parse(text=x))))))
## loop through list
for( i in seq(along=frames) ) {

	#don't need next line, but good to remember if need to exclude a df
	#if (frames[i] != "junk") {

	df <- eval(parse(text=frames[i]))

	for (c in 1:ncol(df)){
		if (is.factor(df[,c])) {
			df[,c] <- gsub("\r\n","",as.vector(df[,c]))
			}
		}

	names(df) <- frames[i]

	}#end first if

}#end first for
#END

Unfortunately, the "names(df)<-frames[i]" line just changes the
name
of the first column in df. How can I overwrite the existing data frame
using the right name?