R，rowMean按data.frame中的列

Question

I have a csv that I've read in and is now a data.frame within R. My file name is MyRData007. My header information goes until row 5 (main column headers are on row 4). My ID is in column A. I simply need to create two separate rowMeans for each ID. The data is in rows 5-147. For the first mean it's columns 4-15; the second mean it's for columns 6-21. Ultimately I'll should have a new variable with a mean for each of the 143 rows. This is what I tried:

> mRNA<-rowMeans(MyRData007)[5:147,(4:15)]
> Protein<-rowMeans(MyRData007)[5:147,(16:21)]

But I get an error?

Error in rowMeans(MyRData007) : 'x' must be numeric

Answer 1

df <- read.table(text='this is a header
                 this is another header
                 this too is one
                 and this is also
                 id code status value
                 1 2 3 4
                 2 32 43 23
                 3 3 43 32
                 4 232 323 55')
df
    V1   V2      V3     V4
1 this   is       a header
2 this   is another header
3 this  too      is    one
4  and this      is   also
5   id code  status  value
6    1    2       3      4
7    2   32      43     23
8    3    3      43     32
9    4  232     323     55

So when you try to call rowMeans you get an error:

rowMeans(df)
Error in rowMeans(df) : 'x' must be numeric

You get this error because you are trying to get the mean of non-numeric values, which makes no sense. Your attempts to subset the data didn't work because you put the brackets outside the call to rowMeans , which tells it to subset the output of rowMeans , not the data going in.

The fundamental problem is that you can't have headers information in a R data.frame . All data in a column of a data frame must be the same type, so if you have characters in some rows, you cant have numbers in others.

How can you fix this?

Read in your data with read.table using the skip = 4 argument. This will make it skip over the header information rows to generate a data frame with only your data. If your file is a .csv you'll also need to specify sep=',' and header=T :

df2 <- read.table(text='this is a header
                 this is another header
                 this too is one
                 and this is also
                 id code status value
                 1 2 3 4
                 2 32 43 23
                 3 3 43 32
                 4 232 323 55', skip = 4, header = T)
rowMeans(df2)
[1]   2.50  25.00  20.25 153.50

read.csv is just a wrapper for read.table and using it is the same as using read.table with the following options:

read.table(file, header = TRUE, sep = ",", fill = TRUE)

Generally, it's better to use read.table since it gives you more control. The most important example being to set stringsAsFactors = FALSE to prevent strings from being converted to factors (an extremely annoying default).