R help - Mar 2010 - Converting "factors" to "numeric" in a dataframe

I am currently trying to write a program that minimises the amount of work
required for “auditable” qPCR data. At the moment I am using an Excel (.csv)
spreadsheet as source data that has been transposed to the column format
required for R to read. Unfortunately, this means I have* *to manually
confirm the whole data set prior to doing any analysis, which is taking a
considerable amount of time! My idea now is to read the raw data in directly
and get R to do the transformation prior to analysis. The problem I now have
is that, upon transposition, the data are converted to “character” in a
matrix, rather than “factor” and “numeric” in a dataframe. I have succeeded
in changing the matrix to a dataframe (via as.data.frame(object)), but this
then converts all the data to “factor” which I can’t use for my analysis
since, other than the column headings, I need the data to be numeric. I have
tried coercing the data to numeric using the as() and as.numeric() commands,
but this has no effect on the data format. I have no experience in
programming and so am at a loss as to what to do: am I making a basic error
in my programming or missing something essential (or both!)?



I am using R version 2.9.0 at the moment, but this will change as soon as I
have sorted this issue out. Below is the code I have put together, as you
can see it is VERY brief but essential to allow my analysis to proceed:



pcrdata<-read.csv("File_path",header=FALSE)

pcrdata<-as.data.frame(t(pcrdata))

pcrdata[2:51]<-as.numeric(as.character(pcrdata))



Any help would be gratefully appreciated,



Mike Glanville

	[[alternative HTML version deleted]]

Hi!

I don't really understand why you do pcrdata<-as.data.frame(t(pcrdata))
Do you need to transpose the dataset? Because read.csv() creates a 
dataframe already.

Something I found really useful recently is the package xlsReadWrite 
where the function read.xls() has an argument colClasses (read.table() 
and read.csv() have it too, but it never worked fine for me) which would 
allow you to specify the class of each column at reading.

Is this what you were looking for or am I completely wrong?

By the way, you could send us the output from str(pcrdata), I mean just 
after reading in the data

Ivan


Le 3/17/2010 15:23, Michael Glanville a écrit :
> I am currently trying to write a program that minimises the amount of work
> required for "auditable" qPCR data. At the moment I am using an
Excel (.csv)
> spreadsheet as source data that has been transposed to the column format
> required for R to read. Unfortunately, this means I have* *to manually
> confirm the whole data set prior to doing any analysis, which is taking a
> considerable amount of time! My idea now is to read the raw data in
directly
> and get R to do the transformation prior to analysis. The problem I now
have
> is that, upon transposition, the data are converted to
"character" in a
> matrix, rather than "factor" and "numeric" in a
dataframe. I have succeeded
> in changing the matrix to a dataframe (via as.data.frame(object)), but this
> then converts all the data to "factor" which I can't use for
my analysis
> since, other than the column headings, I need the data to be numeric. I
have
> tried coercing the data to numeric using the as() and as.numeric()
commands,
> but this has no effect on the data format. I have no experience in
> programming and so am at a loss as to what to do: am I making a basic error
> in my programming or missing something essential (or both!)?
>
>
>
> I am using R version 2.9.0 at the moment, but this will change as soon as I
> have sorted this issue out. Below is the code I have put together, as you
> can see it is VERY brief but essential to allow my analysis to proceed:
>
>
>
> pcrdata<-read.csv("File_path",header=FALSE)
>
> pcrdata<-as.data.frame(t(pcrdata))
>
> pcrdata[2:51]<-as.numeric(as.character(pcrdata))
>
>
>
> Any help would be gratefully appreciated,
>
>
>
> Mike Glanville
>
> 	[[alternative HTML version deleted]]
>
>    
>
>
> ______________________________________________
> R-help@r-project.org mailing list
> 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.
>    
-- 
Ivan CALANDRA
PhD Student
University of Hamburg
Biozentrum Grindel und Zoologisches Museum
Abt. Säugetiere
Martin-Luther-King-Platz 3
D-20146 Hamburg, GERMANY
+49(0)40 42838 6231
ivan.calandra@uni-hamburg.de

**********
http://www.for771.uni-bonn.de
http://webapp5.rrz.uni-hamburg.de/mammals/eng/mitarbeiter.php


	[[alternative HTML version deleted]]

Hi

r-help-bounces at r-project.org napsal dne 17.03.2010 15:23:34:
> I am currently trying to write a program that minimises the amount of 
work
> required for ?auditable? qPCR data. At the moment I am using an Excel 
(.csv)
> spreadsheet as source data that has been transposed to the column format
> required for R to read. Unfortunately, this means I have* *to manually
> confirm the whole data set prior to doing any analysis, which is taking 
a
> considerable amount of time! My idea now is to read the raw data in 
directly
> and get R to do the transformation prior to analysis. The problem I now 
have
> is that, upon transposition, the data are converted to ?character? in a
> matrix, rather than ?factor? and ?numeric? in a dataframe. I have 
succeeded
> in changing the matrix to a dataframe (via as.data.frame(object)), but 
this
> then converts all the data to ?factor? which I can?t use for my analysis
> since, other than the column headings, I need the data to be numeric. I 
have
> tried coercing the data to numeric using the as() and as.numeric() 
commands,
> but this has no effect on the data format. I have no experience in
> programming and so am at a loss as to what to do: am I making a basic 
error
> in my programming or missing something essential (or both!)?
> 
> 
> 
> I am using R version 2.9.0 at the moment, but this will change as soon 
as I
> have sorted this issue out. Below is the code I have put together, as 
you
> can see it is VERY brief but essential to allow my analysis to proceed:
> 
> 
> 
> pcrdata<-read.csv("File_path",header=FALSE)
^^^^^^^^^^^^^^^^

This is supposed to be data frame already. As you did not show us any of 
possible clues of data type like

str(pcrdata)

it is difficult to say.

However from your description your original data are in columns which have 
numeric and character data together which is not possible. I believe that 
there are options for reading such data.
> 
> pcrdata<-as.data.frame(t(pcrdata))
OK. Here you say you get data in columns but they are all character.
> 
> pcrdata[2:51]<-as.numeric(as.character(pcrdata))
> 
Here it depends whether they are all numeric or if some of them shall be 
character (factor). Functions like those above can not be used directly on 
data frames. You need to use apply.

apply(pcrdata, 1, as.character)

Exact sequence of required functions is impossible to guess without 
knowing structure of your objects.

You shall also consult R intro and R data manuals.

Regards
Petr

> 
> 
> Any help would be gratefully appreciated,
> 
> 
> 
> Mike Glanville
> 
>    [[alternative HTML version deleted]]
> 
> ______________________________________________
> R-help at r-project.org mailing list
> 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.

Hi

as you did not provide data it is hard to say what is wrong.

You can see that it is working on dates similar what you described.

 test <- data.frame(x=letters[1:10], y=rnorm(10), z=runif(10))
 test
   x           y          z
1  a  0.09980806 0.32211567
2  b  0.70559139 0.32204076
3  c -1.50514354 0.28267338
............

testnum<-data.frame(t(test[,-1]))
names(testnum)<-test[,1]
testnum
           a         b          c         d         e           f g
y 0.09980806 0.7055914 -1.5051435 0.4421971 0.1041789 -1.54683799 
0.3405809
z 0.32211567 0.3220408  0.2826734 0.8819248 0.5189688  0.05171076 
0.4583101
          h          i         j
y 2.3137394 -0.4953507 0.7668954
z 0.7515886  0.5876854 0.4192073


Regards
Petr



Michael Glanville <michael.glanville19 at googlemail.com> napsal dne 
18.03.2010 12:37:21:
> Thanks Petr, your suggestion has worked to a certain extent. The only 
issue is
> that the sample names don't appear in the final dataframe. However, I 
will 
> persevere and see what I can do.
>  
> Many thanks for you invaluable help,
>  
> Mike
> On 18 March 2010 11:23, Petr PIKAL <petr.pikal at precheza.cz> wrote:
> Hi Michael
> 
> 
> r-help-bounces at r-project.org napsal dne 18.03.2010 12:02:19:
> 
> > Hi petr,
> >
> > Thanks for the reply.
> >
> > My original data is in "comma separated variable" (csv)
format with
> variable
> > names in column 1 and numeric data in the remaining columns. The
> "read.csv"
> > command reads this data set into object name "pcrdata" as a
dataframe
> where
> > the variable names and numeric data are conserved (as required).
> However,
> So why not transpose only numeric part, then set it to data frame and 
add
> column names from first column.
> 
> Something like (untested)
> 
> pcrdata<-read.csv("File_path",header=FALSE)
> numdata<-t(pcrdata[,-1])
> numdata<-data.frame(numdata)
> names(numdata) <- pcrdata[,1]
> 
> Regards
> Petr
>