Converting values to string

Question

I have a data frame

> head(custom.cn.data)
      Gene       Sample_name         CN
1: DDX11L1          F02         0.6788844
2: DDX11L1          C02        -0.9924524
3: DDX11L1          A07        -2.3833559
> 

I want to convert values in CN column < 0 to Del and >0 to Amp

But I am getting error

> a %>% mutate(q = replace(CN, which(CN<0), Del))
Error in replace(CN, which(CN < 0), Del) : object 'Del' not found

> a$CN <- replace(a$CN, which(a$CN < 0), Del)
Error in replace(a$CN, which(a$CN < 0), Del) : object 'Del' not found

Any help?

Answer 1

custom.cn.data$new_column <- ifelse(custom.cn.data$CN < 0, "Del", "Amp")

In this scenario CN == 0 would be Amp

Answer 2

The following code would create a new column with the desired values, 0 values will have NA in this new column.

custom.cn.data[custom.cn.data$CN < 0, "CN_string"] <- "Del" 
custom.cn.data[custom.cn.data$CN > 0, "CN_string"] <- "Amp"

Answer 3

Try to stay within one package (paradigm) for one problem. This will make the code more readable, easier to debug.

At the moment you have data as a data.table object from data.table package, then trying to add a new column using mutate from dplyr package, and within mutate using base replace function.

Solutions:

data.table

# update by row index
custom.cn.data[ CN < 0, NewColumn := "Del" ]
custom.cn.data[ CN >= 0, NewColumn := "Amp" ]

or use ifelse

custom.cn.data[, NewColumn := ifelse(CN < 0, "Del", "Amp") ]

base

see @DavyCat's and @haci's answers.

dplyr

Note, this will change the data.table into data.frame.

# ifelse
custom.cn.data <- custom.cn.data %>%
  mutate(NewColumn = ifelse(CN < 0, "Del", "Amp"))

or use case_when

custom.cn.data <- custom.cn.data %>%
  mutate(NewColumn = case_when(
    CN < 0 ~ "Del",
    CN >= 0 ~ "Amp"))