[R] Manipulations of a data frame or list -- rename or re-assign?

[Jim Lemon]

On 08/03/2013 09:27 PM, Ben Harrison wrote:
> If I am to do a lot of pre-processing operations on some data,
> comprising many steps such as:
>
> -- despike a set of signals
> -- smooth a set of signals
> -- filter
> -- subset
> -- impute missing values
>
> etc... Should I be assigning the result of each operation to a new
> object, and then destroying the old one, or is it fine to reassign the
> object to itself?
>
> df<- sapply(df, someFunction, arg=10)
> OR
> df.new<- sapply(df, someFunction, arg=10)
>
Hi Ben,
If you have read the initial data in from a file you can just use the 
original name. If you don't have another copy of the initial data, it is 
best to save it (e.g. write.table, save) before overwriting the object. 
Fundamentally, don't destroy an object that you can't easily recreate.

Jim