CODEWITHSUNDEEP
rdataframetidyversedplyr
abichat
abichat

Reputation: 2416

Mutate repeats first row value

I have a dataset with taxonomy assignment and I want to extract the genus in a new column.

library(tidyverse)
library(magrittr)
library(stringr)


df <- structure(list(C043 = c(18361L, 59646L, 27575L, 163L, 863L, 3319L, 
                              0L, 6L), C057 = c(20020L, 97610L, 13427L, 1L, 161L, 237L, 2L, 
                                                105L), taxonomy = structure(c(3L, 2L, 1L, 6L, 4L, 4L, 5L, 2L), .Label = c("k__Bacteria;p__Proteobacteria;c__Gammaproteobacteria;o__Enterobacteriales;f__Enterobacteriaceae;g__Enterobacter;NA", 
                                                                                                                          "k__Bacteria;p__Proteobacteria;c__Gammaproteobacteria;o__Enterobacteriales;f__Enterobacteriaceae;g__Enterobacter;s__cloacae", 
                                                                                                                          "k__Bacteria;p__Proteobacteria;c__Gammaproteobacteria;o__Enterobacteriales;f__Enterobacteriaceae;g__Escherichia;s__coli", 
                                                                                                                          "k__Bacteria;p__Proteobacteria;c__Gammaproteobacteria;o__Enterobacteriales;f__Enterobacteriaceae;g__Klebsiella;s__", 
                                                                                                                          "k__Bacteria;p__Proteobacteria;c__Gammaproteobacteria;o__Pseudomonadales;f__Pseudomonadaceae;g__Pseudomonas;s__", 
                                                                                                                          "k__Bacteria;p__Proteobacteria;c__Gammaproteobacteria;o__Pseudomonadales;f__Pseudomonadaceae;g__Pseudomonas;s__stutzeri"
                                                ), class = "factor")), .Names = c("C043", "C057", "taxonomy"), row.names = c(1L, 
                                                                                                                             2L, 3L, 4L, 5L, 6L, 8L, 10L), class = "data.frame")

So this is my function (it works)

extract_genus <- function(str){
  genus <- str_split(str, pattern = ";")[[1]][6] 
  genus %<>% str_sub(start = 4) #%>% as.character
  return(genus)
}

But when I applied it in mutate (with or without as.character), it repeats first row value in the new column. 

df %>% mutate(genus = extract_genus(taxonomy))

   C043  C057                                                                                                                   taxonomy       genus
1 18361 20020     k__Bacteria;p__Proteobacteria;c__Gammaproteobacteria;o__Enterobacteriales;f__Enterobacteriaceae;g__Escherichia;s__coli Escherichia
2 59646 97610 k__Bacteria;p__Proteobacteria;c__Gammaproteobacteria;o__Enterobacteriales;f__Enterobacteriaceae;g__Enterobacter;s__cloacae Escherichia
3 27575 13427         k__Bacteria;p__Proteobacteria;c__Gammaproteobacteria;o__Enterobacteriales;f__Enterobacteriaceae;g__Enterobacter;NA Escherichia
4   163     1     k__Bacteria;p__Proteobacteria;c__Gammaproteobacteria;o__Pseudomonadales;f__Pseudomonadaceae;g__Pseudomonas;s__stutzeri Escherichia
5   863   161          k__Bacteria;p__Proteobacteria;c__Gammaproteobacteria;o__Enterobacteriales;f__Enterobacteriaceae;g__Klebsiella;s__ Escherichia

When I use sapply (but I don't want to, I want a solution with dplyr pipeline), it works.

df_group_gen$genus <- sapply(df_group_gen$taxonomy, extract_genus)

   C043  C057                                                                                                                   taxonomy        genus
1 18361 20020     k__Bacteria;p__Proteobacteria;c__Gammaproteobacteria;o__Enterobacteriales;f__Enterobacteriaceae;g__Escherichia;s__coli  Escherichia
2 59646 97610 k__Bacteria;p__Proteobacteria;c__Gammaproteobacteria;o__Enterobacteriales;f__Enterobacteriaceae;g__Enterobacter;s__cloacae Enterobacter
3 27575 13427         k__Bacteria;p__Proteobacteria;c__Gammaproteobacteria;o__Enterobacteriales;f__Enterobacteriaceae;g__Enterobacter;NA Enterobacter
4   163     1     k__Bacteria;p__Proteobacteria;c__Gammaproteobacteria;o__Pseudomonadales;f__Pseudomonadaceae;g__Pseudomonas;s__stutzeri  Pseudomonas
5   863   161          k__Bacteria;p__Proteobacteria;c__Gammaproteobacteria;o__Enterobacteriales;f__Enterobacteriaceae;g__Klebsiella;s__   Klebsiella

Why mutate doesn't compute as we can expect? I find this question but no answer is provided, only a had hoc code.

Thank you :)

Upvotes: 2

Views: 1468

Answers (1)

Adam Quek
Adam Quek

Reputation: 7153

You can Vectorize your function to allow mutate to occur on every row:

ex_gen <- Vectorize(extract_genus, vectorize.args='str')

df %>% mutate(genus=ex_gen(taxonomy))

Alternatively, you can use rowwise to mutate each row: 

df %>% 
    rowwise() %>% 
    mutate(genus = extract_genus(taxonomy))

Upvotes: 3

Related Questions