R help - Jan 2015 - Passing a Data Frame Name as a Variable in a Function

On 29 Jan 2015, at 07:34 , Jeff Newmiller <jdnewmil at dcn.davis.ca.us>
wrote:
> This approach is fraught with dangers.
> 
> I recommend that you put all of those data frames into a list and have your
function accept the list and the name and use the list indexing operator
mylist[[DFName]] to refer to it. Having functions that go fishing around in the
global environment will be hard to maintain at best, and buggy at worst.
Agreed. However, just to help understand the issue: 
 
DFName is a length-one vector of character strings, not the object that has the
name contained in the string. I.e. you can do nchar(DFName) and presumably get
the value 7, but there is no operation on df.1001 that can tell you the length
of its name. You can (but shouldn't, as per Jeff's note) get from name
to object using get()/assign() and also via some concoctions involving
combinations of eval(), parse(), as.name(), and substitute().

-- 
Peter Dalgaard, Professor,
Center for Statistics, Copenhagen Business School
Solbjerg Plads 3, 2000 Frederiksberg, Denmark
Phone: (+45)38153501
Email: pd.mes at cbs.dk  Priv: PDalgd at gmail.com

Much thanks to everyone for their recommendations!  I agree that fishing in the
global environment isn't ideal & only shows my budding understanding of
R.

For now, I will adapt Chel Hee's
"length(eval(parse(text=DFName))[,1])" solution then fully explore
Jeff's suggestion to put the data frames into a list.

At best, I am implementing an ineloquent approach to:

(1) Add a column to each data frame with a string that is parsed from the
appendage of the data frame name, i.e., string is "1001" from data
frame object of "df.1001"; then,
(2) Bind the rows of all the files.

Any more feedback will be much appreciated!

Cheers,
Alan

On Jan 29, 2015, at 3:11 AM, peter dalgaard wrote:
> 
> On 29 Jan 2015, at 07:34 , Jeff Newmiller <jdnewmil at
dcn.davis.ca.us> wrote:
> 
>> This approach is fraught with dangers.
>> 
>> I recommend that you put all of those data frames into a list and have
your function accept the list and the name and use the list indexing operator
mylist[[DFName]] to refer to it. Having functions that go fishing around in the
global environment will be hard to maintain at best, and buggy at worst.
> 
> Agreed. However, just to help understand the issue: 
> 
> DFName is a length-one vector of character strings, not the object that has
the name contained in the string. I.e. you can do nchar(DFName) and presumably
get the value 7, but there is no operation on df.1001 that can tell you the
length of its name. You can (but shouldn't, as per Jeff's note) get from
name to object using get()/assign() and also via some concoctions involving
combinations of eval(), parse(), as.name(), and substitute().
> 
> -- 
> Peter Dalgaard, Professor,
> Center for Statistics, Copenhagen Business School
> Solbjerg Plads 3, 2000 Frederiksberg, Denmark
> Phone: (+45)38153501
> Email: pd.mes at cbs.dk  Priv: PDalgd at gmail.com
> 
> 
> 
> 
> 
> 
> 
>

On Thu, Jan 29, 2015 at 11:43 AM, Alan Yong <alanyong at caltech.edu>
wrote:
> Much thanks to everyone for their recommendations!  I agree that fishing in
the global environment isn't ideal & only shows my budding understanding
of R.
>
> For now, I will adapt Chel Hee's
"length(eval(parse(text=DFName))[,1])" solution then fully explore
Jeff's suggestion to put the data frames into a list.
If you have to go down this route, at least do nrow(get(DFName))
> (1) Add a column to each data frame with a string that is parsed from the
appendage of the data frame name, i.e., string is "1001" from data
frame object of "df.1001"; then,
> (2) Bind the rows of all the files.
I'd highly recommend learning a little functional programming such as
the use of lapply (e.g. http://adv-r.had.co.nz/Functionals.html).
Then you can easily do:

csvs <- dir(pattern = "\\.csv$")
all <- lapply(csvs, read.csv)
one <- do.call("rbind", all)

to find all the csv files in a directory, load into a list and then
collapse into a single data frame.

You're much better off learning how to do this than futzing around
with named objects in the global environment.

Hadley

-- 
http://had.co.nz/