Using gsub on columns in R

Question

I have a data frame (data) in R with thousands of rows and 10 columns. 9 of the columns contain factors with several levels.

Here is a small portion of the data frame.

A gr1

10 303.90

11 304.1

12 303.6

13 303.90 obs

14 303.90k

As an example, one factor has a level that is "303.90" and another level that is "303.90 obs". I want to change the "303.90 obs" to "303.90". I am using the following command to edit the names of the level.

data[] = as.data.frame(lapply(data, function(x) {x = gsub("303.90 obs","303.90", fixed = T, x)}))

But this is not changing the level "303.90 obs" to "303.90". It just stays the same. Still this command works for other strings, eg. "303.9" gets changed to "303.90" when I use:

data[] = as.data.frame(lapply(data, function(x) {x = gsub("303.9 obs","303.90", fixed = T, x)}))

Any suggestions to why this might be ?

First, remove `x=` from the `lapply`. Second, please provide a small sample of data: http://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example — emilliman5, Mar 22 '18 at 14:34
So you also want that `303.90k` should be changed to `303.90`? Then you need regular expression. — Tobias, Mar 22 '18 at 15:00
I think my code might be replacing the "303.90" part of "303.90 obs" but still printing the rest of the string. Could this be the case ? — gwarr, Mar 22 '18 at 15:35
Yes. 303.90k should also be changed. I will try to use regular expressions, but I thing gsub should work for this example anyway. Need to figure out why it is not working before moving on to regular expressions. — gwarr, Mar 22 '18 at 15:41
I suspect this is a issue with the file I´m importing, because I´m not able to convert this level "303.90 obs". I have tried making an example data.frame from scratch in R, including factors with several levels in similar format "303.90 ..." and I have been able to convert those successfully using all the methods mentioned here. I have to take a closer look at the file I´m importing. Maybe there are some hidden characters ? — gwarr, Mar 23 '18 at 11:46

Tobias · Answer 1 · 2018-03-23T07:24:04.133

I'm not that familiar with lapply therefore my solution simply loops over the columns of the dataframe. This works as it should.

col1 <- 1:10
col2 <- 21:30
col3 <- c("503.90", "303.90 obs", "803.90sfsdf sf", "203.90 obs", "303.90", "103.90 obs", "303.90", "403.90 obs", "803.90sfsdf sf", "303.90 obs")
col4 <- c("303.90", "303.90 obs", "303.90", "203.90 obs", "303.90", "107.40fghfg", "303.90", "303.90 obs", "303.90", "303.90 obs")

data <- data.frame(col1, col2, col3, col4)

data$col3 <- as.factor(data$col3)
data$col4 <- as.factor(data$col4)

for(i in 3:4) {
  matchedExpression = regexpr(pattern = "\\d+\\.\\d+", text = data[,i])
  data[,i] = regmatches(x = data[,i], m = matchedExpression)
  data[,i] <- as.factor(data[,i])
}

EDIT

OP changed description. To change all factors to 303.90regex is a better solution. However, more information are necessary from the OP to give a general solution e.g. is it only 303.90 which should be changed?

EDIT2

Updated the script since OP provided more information e.g. columns can have different factors than 303.90.

This method is no working for me. The "303.90 obs" is not being changed to "303.90". — gwarr, Mar 22 '18 at 15:14
When you copy&paste the script and execute it, you will see that it works. If my code does not work on your dataframe you did not provide enough information. See http://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example — Tobias, Mar 22 '18 at 15:18

Using gsub on columns in R

1 Answers1

Linked