Collectives™ on Stack Overflow

Find centralized, trusted content and collaborate around the technologies you use most.

Teams

Q&A for work

Connect and share knowledge within a single location that is structured and easy to search.

I have a problem using data from a tab delimited data file imported with read.delim .

Most of the columns contain numerical data which I need to do a t.test for. Unfortunately I always get this error:

Error in if (stderr < 10 * .Machine$double.eps * max(abs(mx), abs(my))) 
            stop("data are essentiallyconstant") :
missing value where TRUE/FALSE needed
In addition: Warning messages:
1: In mean.default(x) : argument is not numeric or logical: returning NA
2: In mean.default(y) : argument is not numeric or logical: returning NA
I noticed that this only happens with vectors that consist of different levels.
It won't even perform simple numerical operations like vector[1] + vector[2] for leveled vectors.
Vectors without levels work fine, though.
How can I use the data in the leveled vectors for calculation?
Thank you
                Reproducible example on toy data, please. Also, if you could, try formatting your code using the '1010101' button.
– Dirk Eddelbuettel
                Jul 14, 2010 at 14:36
I have been able to reproduce your error message with the following small example:
x = as.factor(1:5)
y = as.factor(1:5)
t.test(x, y)
yields
Error in if (stderr < 10 * .Machine$double.eps * max(abs(mx), abs(my))) stop("data are essentially constant") : 
  missing value where TRUE/FALSE needed
In addition: Warning messages:
1: In mean.default(x) : argument is not numeric or logical: returning NA
2: In mean.default(y) : argument is not numeric or logical: returning NA
The problem is you are trying to perform a t-test on non-numeric vectors. Addition likewise is not defined for factors:
x + y
yields
[1] NA NA NA NA NA
Warning message:
In Ops.factor(x, y) : + not meaningful for factors
The warning gives keen insight as to what is amiss and also explains why your t-test is not working.
To fix the problem, you need to do as ilya suggests: convert your vectors to numeric with as.numeric(as.character())
                The real solution is to find out why your numeric vectors are being converted to factors in the first place.
– hadley
                Jul 15, 2010 at 7:11
                Thanks a lot.  Converting the leveled vectors to numeric with as.numeric(as.character(vector)) solved the problem. Indeed the entries in the leveled vector were factoral.
– jev
                Jul 15, 2010 at 7:48
                Yeah - "leveled vectors" sounds a lot like a factor to me.  On read.delim, you can set the stringsAsFactors argument to FALSE, which will prevent this from happening.  There's also a global stringsAsFactors option (if you look at ?read.delim, you can see the defaults check this option).  I highly recommend it.
– Matt Parker
                Jul 14, 2010 at 15:01
                See also the R FAQ on turning factors to numeric (naked as.numeric doesn't work). cran.r-project.org/doc/FAQ/…
– Richie Cotton
                Jul 14, 2010 at 15:05
You say "Most of the columns contain numerical data". That's the problem. Only when all columns contain numerical data, can the function apply used without changing the data type. If there is any non-numerical data in other columns, you should change the data type in the function apply:
        pvalue<-apply(x,1,ttest<-function(tmp { 
                              if(length(unique(c(tmp[5],tmp[7],tmp[9])))!=1 && 
                              length(unique(c(tmp[11],tmp[13],tmp[15])))!=1) 
                              t.test(c(as.numeric(tmp[5]),as.numeric(tmp[7]),
                              as.numeric(tmp[9])), c(as.numeric(tmp[11]), 
                              as.numeric(tmp[13]),as.numeric(tmp[15])))$p.value 
                              else NA})
        
Thanks for contributing an answer to Stack Overflow!
Please be sure to answer the question. Provide details and share your research!
But avoid …
Asking for help, clarification, or responding to other answers.
Making statements based on opinion; back them up with references or personal experience.
To learn more, see our tips on writing great answers.