R help - Nov 2016 - Different results when converting a matrix to a data.frame

Hi All,

I build an empty dataframe to fill it will values later. I did the 
following:

-- cut --
matrix(NA, 2, 2)
     [,1] [,2]
[1,]   NA   NA
[2,]   NA   NA
> data.frame(matrix(NA, 2, 2))
  X1 X2
1 NA NA
2 NA NA
> as.data.frame(matrix(NA, 2, 2))
  V1 V2
1 NA NA
2 NA NA
-- cut --

Why does data.frame deliver different results than as.data.frame with 
regard to the variable names (V instead of X)?

Kind regards

Georg

	[[alternative HTML version deleted]]

I will start by admitting I don't know the answer to your question.

However, I am responding because I think this should not be an issue in real
life use of R. Data frames are lists of distinct vectors, each of which has its
own reason for being present in the data, and normally each has its own storage
mode. Your use of a matrix as a short cut way to create many columns at once
does not change this fundamental difference between data frames and matrices.
You should not be surprised that putting the finishing touches on this
transformation takes some personal attention.

Normally you should give explicit names to each column using the argument names
in the data.frame function. When using a matrix as a shortcut, you should either
immediately follow the creation of the data frame with a names(DF)<-
assignment, or wrap it in a setNames function call.

setNames( data.frame(matrix(NA, 2, 2)), c( "ColA", "ColB" )
)

Note that using a matrix to create many columns is memory inefficient, because
you start by setting aside a single block of memory (the matrix) and then you
move that data column at a time to separate vectors for use in the data frame.
If working with large data you might want to consider allocating each column
separately from the beginning.

N <- 2
nms <- c( "A", "B" )
as.data.frame( setNames( lapply( nms, function(n){ rep( NA, 2 ) } ), nms ) )

which is not as convenient, but illustrates that data frames are truly different
than matrices.
-- 
Sent from my phone. Please excuse my brevity.

On November 16, 2016 7:20:38 AM PST, G.Maubach at weinwolf.de
wrote:
>Hi All,
>
>I build an empty dataframe to fill it will values later. I did the 
>following:
>
>-- cut --
>matrix(NA, 2, 2)
>     [,1] [,2]
>[1,]   NA   NA
>[2,]   NA   NA
>> data.frame(matrix(NA, 2, 2))
>  X1 X2
>1 NA NA
>2 NA NA
>> as.data.frame(matrix(NA, 2, 2))
>  V1 V2
>1 NA NA
>2 NA NA
>-- cut --
>
>Why does data.frame deliver different results than as.data.frame with 
>regard to the variable names (V instead of X)?
>
>Kind regards
>
>Georg
>
>	[[alternative HTML version deleted]]
>
>______________________________________________
>R-help at r-project.org mailing list -- To UNSUBSCRIBE and more, see
>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 Nov 16, 2016, at 8:43 AM, Jeff Newmiller <jdnewmil at
dcn.davis.ca.us> wrote:
> 
> I will start by admitting I don't know the answer to your question.
> 
> However, I am responding because I think this should not be an issue in
real life use of R. Data frames are lists of distinct vectors, each of which has
its own reason for being present in the data, and normally each has its own
storage mode. Your use of a matrix as a short cut way to create many columns at
once does not change this fundamental difference between data frames and
matrices. You should not be surprised that putting the finishing touches on this
transformation takes some personal attention.
> 
> Normally you should give explicit names to each column using the argument
names in the data.frame function. When using a matrix as a shortcut, you should
either immediately follow the creation of the data frame with a names(DF)<-
assignment, or wrap it in a setNames function call.
> 
> setNames( data.frame(matrix(NA, 2, 2)), c( "ColA",
"ColB" ) )
> 
> Note that using a matrix to create many columns is memory inefficient,
because you start by setting aside a single block of memory (the matrix) and
then you move that data column at a time to separate vectors for use in the data
frame. If working with large data you might want to consider allocating each
column separately from the beginning.
> 
> N <- 2
> nms <- c( "A", "B" )
> as.data.frame( setNames( lapply( nms, function(n){ rep( NA, 2 ) } ), nms )
)
> 
> which is not as convenient, but illustrates that data frames are truly
different than matrices.
> -- 
> Sent from my phone. Please excuse my brevity.
> 
> On November 16, 2016 7:20:38 AM PST, G.Maubach at weinwolf.de wrote:
>> Hi All,
>> 
>> I build an empty dataframe to fill it will values later. I did the 
>> following:
>> 
>> -- cut --
>> matrix(NA, 2, 2)
>>    [,1] [,2]
>> [1,]   NA   NA
>> [2,]   NA   NA
>>> data.frame(matrix(NA, 2, 2))
>> X1 X2
>> 1 NA NA
>> 2 NA NA
>>> as.data.frame(matrix(NA, 2, 2))
>> V1 V2
>> 1 NA NA
>> 2 NA NA
>> -- cut --
>> 
>> Why does data.frame deliver different results than as.data.frame with 
>> regard to the variable names (V instead of X)?
They are two different functions:

It's fairly easy to see by looking at the code:

as.data.frame.matrix uses: names(value) <- paste0("V", ic)  when
there are no column names and data.frame calls make.names which prepends an
"X" as the first letter of invalid or missing names.


As to why the authors did it this way, I'm unable to comment.
>> 
>> Kind regards
>> 
>> Georg
>> 
>> 	[[alternative HTML version deleted]]

David Winsemius
Alameda, CA, USA