Mutate and replace question

[arozebo]

Hi, I am using mutate and replace to change something in a data frame. In the preview of the file in the notepad it shows me that what I want to be replaced is indeed replaced with the right thing but when I "View" the file or export it and open it with excel, it is not actually replaced. It is still the original thing.

Here is the code I am using to convert the file to a data frame (I imported it as .txt file)

Convert to data frame and include _df at end of file name.

Merge.CD8.Phenotype.Part.2_cell_seg_data_df <- data.frame(Merge.CD8.Phenotype.Part.2_cell_seg_data)

Mutate CD8 to CD8+

Merge.CD8.Phenotype.Part.2_cell_seg_data_df %>% 
  mutate(Phenotype = as.character(Phenotype)) %>% 
  mutate(Phenotype = replace(Phenotype, Phenotype == 'CD8', 'CD8+'))

Phenotype Cell.ID Total.Cells

Other 91 NA
CD8+ 92 NA
Other 93 NA

The column header is "Phenotype" and I want to change all the CD8 to CD8+.  As I mentioned above, in the preview window in the notepad, it says CD8+, but when I "View" it or export it, it still says CD8.

Can anyone help me figure out what I am doing wrong?  Do you need more information, sorry, I am new to R.

Thanks for your help,
Aaron

[martin.R]

Try this:

Merge.CD8.Phenotype.Part.2_cell_seg_data_df <- 
  Merge.CD8.Phenotype.Part.2_cell_seg_data_df %>% 
  mutate(Phenotype = as.character(Phenotype)) %>% 
  mutate(Phenotype = ifelse(Phenotype == 'CD8', 'CD8+', Phenotype))

You need to save the result to another variable or itself (as in this example) if you want to use the result further in your analysis.

[arozebo]

Awesome! That did it.
Can you explain to me how the "ifelse" statement works there to fix the issue and why you had to include the additional Phenotype after the CD8+? I just want a better understanding of how this works to help me learn R.

Thanks again, I appreciate your help,
Aaron

[martin.R]

The ifelse() evaluates the condition Phenotype == 'CD8': if it's TRUE, then it returns 'CD8+', else it returns the existing value of Phenotype, so this replaces the value as intended.

In base R you could replace the value by subsetting the condition, but the above is the most straightforward way within dplyr::mutate().

[dromano]

Just short note to add to @martin.R's explanation: Your original code

mutate(Phenotype = replace(Phenotype, Phenotype == 'CD8', 'CD8+'))

mixes the requirements of replace(), which expects a vector of indices where the replacement should occur, and the requirements of ifelse() (or if_else(), an updated version), which expects a logical vector to indicate when replacement should occur. With this in mind, the following modification of your replace() command will work:

mutate(Phenotype = replace(Phenotype, which(Phenotype == 'CD8'), 'CD8+'))

[arozebo]

Thanks, that is very helpful for my understanding!

[arozebo]

And thanks to you too, dromano, that also helps me think about the logic behind the code better!