Recieving incorrect column name after adding a new column to dataframe

Question

I have a file titled "femalet3" which has a column consisting of string values. All the other columns have a numerical value, which is why I use the code below:

femalet3$mean.f<-data.frame(mean.f=femalet3[,1], mean.f=rowMeans(femalet3[,-1]))

The idea for this came from: Calculate row means on subset of columns

The issue is that when I run this line I receive this output:

Significance GSM1311846 GSM1311847 mean.f.mean.f mean.f.mean.f.1
Vsig         88.35497   83.16820          VSig        85.40076

The issue is that I have several "mean.f" and the value for "Significance" is copied over into mean.f column. I did colnames(femalet3) and the output is:

"Significance" "GSM1311840" "GSM1311841" "GSM1311842" "GSM1311843"        "GSM1311844" "GSM1311845"  "GSM1311846"  "GSM1311847"  "mean.f"

There is apparently only one "mean.f" despite the output earlier. I don't think I am using the line of code taken from the other Q&A correctly and it may be causing this error in formatting. The desired output is:

          Significance GSM1311846 GSM1311847 mean.f
          Vsig         88.35497   83.16820    85.40076

Welcome to the SO! We could help you better if you provided an example that produces the error (MVCE). If that is not feasible, is your file loaded into R as a `data.frame` or a `matrix`? Try `str(femalet3)` to check that each column is loaded as you would expect. — nya, Nov 25 '19 at 15:31
Hello, thank you for the response. The file is a data.frame and after doing str(femalet3) is says that there are two variables in mean.f, which is not what I want : $ mean.f : Factor w/ 3 levels "Sig","VSig","VVSig": 2 1 2 1 1 1 1 1 1 1 ... $ mean.f.1: num 85.4 77.2 87.3 84.7 84.3 ... — Gil Ong, Nov 25 '19 at 15:47
Hey Gil, ok you should not force a data.frame into a column of a data.frame.. First do, results = data.frame(mean.f=femalet3[,1], mean.f=rowMeans(femalet3[,-1])). head(results). Is this what you want? — StupidWolf, Nov 25 '19 at 16:05
@StupidWolf I tried the command but it is still splitting mean.f into two variables, which are "mean.f" and "mean.f.1". My goal is to have only the mean value for mean.f . — Gil Ong, Nov 25 '19 at 22:16
Sorry I made a typo, so it should be results=data.frame(var=femalet3[,1], mean.f=rowMeans(femalet3[,-1])) — StupidWolf, Nov 25 '19 at 22:19
The first column called var, is the id of the row variable, the 2nd column gives u the rowMeans on subset of columns — StupidWolf, Nov 25 '19 at 22:20
@StupidWolf it seems to be working, let me run the whole thing! — Gil Ong, Nov 25 '19 at 22:22
@StupidWolf it's working properly now, I suppose it was because I was calling the old and new columns the same name? Thank you very much for your help! — Gil Ong, Nov 25 '19 at 22:32
No not really, you were trying to force a data.frame into a column of a data frame. Ok below I write a solution and explain why you encountered a problem. — StupidWolf, Nov 25 '19 at 22:36

score 1 · Answer 1 · answered Nov 25 '19 at 22:44

You encountered a problem from this line of code:

femalet3$mean.f<-data.frame(mean.f=femalet3[,1], mean.f=rowMeans(femalet3[,-1]))

femalet3 is a data.frame to start with. If you try to assign another data frame to a column, it gives you something with a weird structure.

I simulate your dataset below to show where the error occurs:

femalet3 <- data.frame(Significance = letters[1:10],matrix(rnorm(80),ncol=8))
colnames(femalet3)[-1] = c("GSM1311840","GSM1311841","GSM1311842",
"GSM1311843","GSM1311844","GSM1311845","GSM1311846","GSM1311847")
femalet3$mean.f<-data.frame(mean.f=femalet3[,1], mean.f=rowMeans(femalet3[,-1]))

head(femalet3)
  Significance  GSM1311840 GSM1311841  GSM1311842  GSM1311843 GSM1311844
1            a -0.09282641  0.0753268 -0.04400652  0.02442526  0.3065423
2            b  1.14718259  0.6062297 -0.08556210  0.15121682  1.6412273
3            c -1.45645947 -1.6808505 -1.93452662 -0.06121562  1.9080640
4            d  0.03955011  1.5496713 -0.27779819 -0.69083631  0.8331726
5            e -0.61881124  1.2798835 -0.55046474 -0.61394703  2.3530607
6            f  1.77918616  0.5156059  0.37311045  1.77081855 -0.8689152
  GSM1311845 GSM1311846 GSM1311847 mean.f.mean.f mean.f.mean.f.1
1  1.1210784  0.6891616  0.7314997             a       0.3514002
2  1.8341236  3.0722572  0.9026674             b       1.1586678
3 -0.5721591  2.8964295 -2.0082267             c      -0.3636181
4  1.1212192  0.2129126  0.9595494             d       0.4684301
5 -0.6253303  1.0512457 -1.2166623             e       0.1323718
6  0.4963209 -0.5864916  0.4429023             f       0.4903172

This embeds a data.frame within the mean.f column in your data.frame:

ncol(femalet3)
10

head(femalet3$mean.f)
   mean.f   mean.f.1
1       a  0.3514002
2       b  1.1586678
3       c -0.3636181
4       d  0.4684301
5       e  0.1323718

We remove the previous weird column:

femalet3$mean.f <- NULL

To avoid this, what you simply need is:

femalet3$mean.f<-rowMeans(femalet3[,-1])
head(femalet3)

> head(femalet3)
  Significance  GSM1311840 GSM1311841  GSM1311842  GSM1311843 GSM1311844
1            a -0.09282641  0.0753268 -0.04400652  0.02442526  0.3065423
2            b  1.14718259  0.6062297 -0.08556210  0.15121682  1.6412273
3            c -1.45645947 -1.6808505 -1.93452662 -0.06121562  1.9080640
4            d  0.03955011  1.5496713 -0.27779819 -0.69083631  0.8331726
5            e -0.61881124  1.2798835 -0.55046474 -0.61394703  2.3530607
6            f  1.77918616  0.5156059  0.37311045  1.77081855 -0.8689152
  GSM1311845 GSM1311846 GSM1311847     mean.f
1  1.1210784  0.6891616  0.7314997  0.3514002
2  1.8341236  3.0722572  0.9026674  1.1586678
3 -0.5721591  2.8964295 -2.0082267 -0.3636181
4  1.1212192  0.2129126  0.9595494  0.4684301
5 -0.6253303  1.0512457 -1.2166623  0.1323718
6  0.4963209 -0.5864916  0.4429023  0.4903172

Recieving incorrect column name after adding a new column to dataframe

1 Answers1