Hello useRs! So, I have a random question. I'm trying to build a character string, then evaluate it. I think an example would be the easiest way to explain: kern.vec = c("rbfdot","polydot") for( j in 1:length( kern.vec ) ) { formula = paste("ksvm( ind ~ . , data=d.temp[,c(ind_col,dep_cols)], kernel =",kern.vec[j],", prob.model=T )") svm = eval( parse( text=formula ) ) ... } The problem I always seem to have is that in the formula, I need to have quotes around "rbfdot" (for example). But, when I paste the expression together, it removes the quotes. Is there a better way to do this (or at least a way around this problem)? My method seems a bit kludgy :) Thanks for all your help! Josh [[alternative HTML version deleted]]
What's the "big picture" of what you're trying to do? eval(parse(text = )) is often a less than optimal idea. Some guesses: Are you trying to construct a formula object (in the strict sense of something that you pass to a modeling function)? Maybe lazy evaluation of the deparse(substitute(x)) flavor might help here? Michael On Sat, May 5, 2012 at 2:42 PM, Josh Browning <rockclimber112358 at gmail.com> wrote:> Hello useRs! > > So, I have a random question. ?I'm trying to build a character string, then > evaluate it. ?I think an example would be the easiest way to explain: > > > ? ?kern.vec = c("rbfdot","polydot") > ? ?for( j in 1:length( kern.vec ) ) > ? ?{ > ? ? ?formula ? ?= paste("ksvm( ind ~ . , > data=d.temp[,c(ind_col,dep_cols)], kernel =",kern.vec[j],", prob.model=T )") > ? ? ?svm ? ?= eval( parse( text=formula ) ) > ? ? ... > ? ?} > > > The problem I always seem to have is that in the formula, I need to have > quotes around "rbfdot" (for example). ?But, when I paste the expression > together, it removes the quotes. ?Is there a better way to do this (or at > least a way around this problem)? ?My method seems a bit kludgy :) > > Thanks for all your help! > > Josh > > ? ? ? ?[[alternative HTML version deleted]] > > ______________________________________________ > R-help at r-project.org mailing list > https://stat.ethz.ch/mailman/listinfo/r-help > PLEASE do read the posting guide http://www.R-project.org/posting-guide.html > and provide commented, minimal, self-contained, reproducible code.
On May 5, 2012, at 2:42 PM, Josh Browning wrote:> Hello useRs! > > So, I have a random question. I'm trying to build a character > string, then > evaluate it.Actually you are trying to build a language object , a call or an expression. You might have gotten further with: do.call(ksvm, list( ... named arguments ...)> I think an example would be the easiest way to explain: > > > kern.vec = c("rbfdot","polydot") > for( j in 1:length( kern.vec ) ) > { > formula = paste("ksvm( ind ~ . , > data=d.temp[,c(ind_col,dep_cols)], kernel =",kern.vec[j],", > prob.model=T )") > svm = eval( parse( text=formula ) ) > ... > } > > > The problem I always seem to have is that in the formula, I need to > have > quotes around "rbfdot" (for example). But, when I paste the > expression > together, it removes the quotes. Is there a better way to do this > (or at > least a way around this problem)? My method seems a bit kludgy :)Perhaps looking at either: ?substitute ?bquote kern.vec = c("rbfdot","polydot") for( j in 1:length( kern.vec ) ) { formula = bquote(expression( ksvm( ind ~ . , data=d.temp[,c(ind_col,dep_cols)], kernel =.(kern.vec[j] ) , prob.model=T ) )) print(formula) } expression(ksvm(ind ~ ., data = d.temp[, c(ind_col, dep_cols)], kernel = "rbfdot", prob.model = T)) expression(ksvm(ind ~ ., data = d.temp[, c(ind_col, dep_cols)], kernel = "polydot", prob.model = T)) Notice that the values for kern.vec are 'character' which is what you passed them (and what you seem to be requesting. If you wanted the values of those named objects you might try get(<name>). (I'm not a user of whatever package has `ksvm` in it, so I'm not aware of whether 'rbfdot' is supposed to be a character value as a parameter or if those are named objects, and I'm not running out to identify the package and then to locate a working example to test the eval-result. Those are details you should have provided.)>-- David Winsemius, MD West Hartford, CT