plyr does not preserve names in nested calls

[karthik]

See related Stack Overflow question will all the details to replicate the problem.

A fix on at the plyr end would be to check for an existing .id column before adding one so it could be renamed appropriately.

[krlmlr]

Now the .id column can be renamed by passing a value to the new .id parameter in ldply.

[hadley]

Maybe it should automatically pick a new column if there's already an .id column?

[krlmlr]

I think that most of the times we want to give an expressive name to the .id column anyway (which motivated my patch). Say, if we have two lists that define size and shape of something and run a nested ldply for all combinations of size and shape, we want two columns Size and Shape in the resulting data frame, and not .id and .id2.