R help - Oct 2008 - convert matrix to dataframe with repeating row names

Hi R,

 

I have a matrix x with repeating row names. 
> dim(x)
[1] 862  19

 

zz<-matrix(0,4,4)

rownames(zz)=c("a","a","b","b")

data.frame(zz) (?)

 

 

I need to use x in a linear regression

lm(as.formula(paste("final_dat[,5]~",paste(colnames(x),collapse="+"))),x
)

this gives me a error

 

Error in model.frame.default(formula
as.formula(paste("final_dat[,5]~",  :

  'data' must be a data.frame, not a matrix or an array

 

 
> sessionInfo()
R version 2.7.1 (2008-06-23) 

i386-pc-mingw32 

 

locale:

LC_COLLATE=English_United States.1252;LC_CTYPE=English_United
States.1252;LC_MONETARY=English_United
States.1252;LC_NUMERIC=C;LC_TIME=English_United States.1252

 

attached base packages:

[1] stats     graphics  grDevices utils     datasets  methods   base


 

other attached packages:

[1] xlsReadWrite_1.3.2

 

Thanks in advance

Ravi

 

 

 

This e-mail may contain confidential and/or privileged i...{{dropped:13}}

> I have a matrix x with repeating row names. 
> zz<-matrix(0,4,4)
> 
> rownames(zz)=c("a","a","b","b")
> 
> data.frame(zz) (?)
The row names on a data frame should be unique.  You can try 
as.data.frame(xx, row.names=FALSE) to convert zz to be a data frame.  If 
you need the row name information, add it as a column in the data frame, 
e.g. mydataframe$rnames <- rownames(zz).  (Note to R-Core: the 
documentation for as.data.frame doesn't mention the usage of 
row.names=FALSE to ignore row names, but it seems to work consistently. 
Does the help page for as.data.frame need updating?)
>
lm(as.formula(paste("final_dat[,5]~",paste(colnames(x),collapse="+"))),x
> )
> 
> this gives me a error
> Error in model.frame.default(formula >
as.formula(paste("final_dat[,5]~",  :
> 
>   'data' must be a data.frame, not a matrix or an array
I suspect that if you try class(x), it will be a matrix, not the requisite 
data frame.

Regards,
Richie.

Mathematical Sciences Unit
HSL


------------------------------------------------------------------------
ATTENTION:

This message contains privileged and confidential inform...{{dropped:20}}

> > The row names on a data frame should be unique.  You can try
> > as.data.frame(xx, row.names=FALSE) to convert zz to be a data frame. 
If
> > you need the row name information, add it as a column in the data 
frame,
> > e.g. mydataframe$rnames <- rownames(zz).  (Note to R-Core: the
> > documentation for as.data.frame doesn't mention the usage of
> > row.names=FALSE to ignore row names, but it seems to work 
consistently.
> > Does the help page for as.data.frame need updating?)
> 
> No.  row.names=FALSE is not intended to work, and did you check every 
> single as.data.frame() method?
> 
> It just so happens that for the matrix method invalid input for 
> 'row.names' results in setting default row names.  Other methods
may
> differ.
row.names=FALSE seems a natural way of supressing existing row names to 
me, since it corresponds nicely to using row.names=FALSE in write.csv. 
Currently it seems that if a matrix has duplicate row names, then 
converting it to be a data frame requires

rnames <- rownames(mymatrix)
rownames(mymatrix) <- NULL
as.data.frame(mymatrix)
rownames(mymatrix) <- rnames 

Ideally, three of these lines of code shouldn't really need to be there.

If you disagree that allowing row.names=FALSE is a good idea, or you don't 
want to change the function interface, then perhaps having as.data.frame 
check for duplicates and throwing a warning (rather than an error) would 
be preferable behaviour.  I do realise that there are dozens of 
as.data.frame methods, and the documentation does state that "Few of the 
methods check for duplicated row names", but it would be beneficial from a 
user standpoint.

Regards,
Richie.

Mathematical Sciences Unit
HSL


------------------------------------------------------------------------
ATTENTION:

This message contains privileged and confidential inform...{{dropped:20}}