select numeric columns and one column specified by name from data frame

Question

I have a data frame which contains both numeric and non-numeric columns, say

df <- data.frame(v1=1:20,v2=1:20,v3=1:20,v4=letters[1:20],v5=letters[1:20])

To select only the non-numeric columns I would use

fixCol <- !sapply(df,is.numeric)

But now I also want to include a specific numeric column, say v2. My data frame is very big and the order of the columns changes, so I cannot index it using a number, I really want to use the name 'v2'. I tried

fixCol$v2 = TRUE

but that gives me the warning In fixCol$FR = TRUE : Coercing LHS to a list which makes it impossible to subset my original data frame to get only fixCol

df[,fixCol]

gives: Error in .subset(x, j) : invalid subscript type 'list'

In the end my goal is to scale all numeric columns of my data frame except this one specified column, using something like this

scaleCol = !fixCol
df_scaled = cbind(df[,fixCol], sapply(df[,scaleCol],scale))

How can I best do this?

akrun · Accepted Answer · 2016-02-12 10:48:42Z

3

We can use a OR condition (|) to get a logical index and then subset the columns of 'df'.

df1 <- df[!sapply(df, is.numeric)|names(df)=='v2']
head(df1,2)
#  v2 v4 v5
#1  1  a  a
#2  2  b  b

answered Feb 12, 2016 at 10:48

akrun

891k38 gold badges590 silver badges700 bronze badges

Sign up to request clarification or add additional context in comments.

2 Comments

Ciska Over a year ago

this works great, thanks! How do I make df2 which contains all columns NOT put in df1, without reversing the whole sapply statement? I tried df2 <- df[,!names(df1)] or df2 <- df[,-c(names(df1))] but both give an error. I need df2 as well in order to be able to scale df2 and put it in a new dataframe together with df1.

akrun Over a year ago

@Ciska We can either use setdiff or %in% i.e. df2 <- df[setdiff(names(df), names(df1))]

Sotos · Accepted Answer · 2016-02-12 10:52:21Z

2

fixCol <- !sapply(df,is.numeric)
fixCol <- df[, fixCol]
fixCol$v2 <- df[colnames(df)=="v2"]
head(fixCol)
 # v4 v5 v2
#1  a  a  1
#2  b  b  2
#3  c  c  3
#4  d  d  4
#5  e  e  5
#6  f  f  6

answered Feb 12, 2016 at 10:52

Sotos

51.6k6 gold badges35 silver badges69 bronze badges

1 Comment

Ciska Over a year ago

thanks, this works! The answer from akrun uses a little less lines but I am happy with both. Thanks again!

Collectives™ on Stack Overflow

select numeric columns and one column specified by name from data frame

2 Answers 2

2 Comments

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

2 Comments

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related