CODEWITHSUNDEEP
rstring
Calosoma
Calosoma

Reputation: 71

Split string in R and reassambling column by column

I have a complex string splitting Problem in R. In my dataframe I have a column with strings of different lengths. 

   Site  Class
   A1    D2.13
   A2     E1.4
   A3     FA.1
   A4    H2.14
   A5        F
   AR       G1

Now I want to add new columns that reassemble the string character by character, while the dot should be "ignored" in character by character. 

   Site Class1 Class2 Class3 Class4
   A1      D     D2   D2.1  D2.13
   A2      E     E1   E1.4     NA
   A3      F     FA   FA.1     NA
   A4      H     H2   H2.1  H2.14
   A5      F     NA     NA     NA
   AR      G     G1     NA     NA

Test data:

structure(list(Site = c("A1", "A2", "A3", "A4", "A5", "AR"), 
           Class = c("D2.13", "E1.4", "FA.1", "H2.14", "F","G1")), 
           class = "data.frame", row.names = c(NA, -6L))

Upvotes: 1

Views: 78

Answers (3)

arg0naut91
arg0naut91

Reputation: 14774

Similar idea to what @Sotos did (crucial parts are Reduce and strsplit) with a bit of different configuration:

library(data.table)

df <- setDT(df)[, .(Class = Reduce(paste0, unlist(strsplit(as.character(Class), split = "")), accumulate = T)), 
                by = Site][
                  !grepl("\\.$", Class)][, nr := paste0("Class", rleid(Class)), by = Site]

dcast(df, Site ~ nr, value.var = "Class")

Output:

   Site Class1 Class2 Class3 Class4
1:   A1      D     D2   D2.1  D2.13
2:   A2      E     E1   E1.4   <NA>
3:   A3      F     FA   FA.1   <NA>
4:   A4      H     H2   H2.1  H2.14
5:   A5      F   <NA>   <NA>   <NA>
6:   AR      G     G1   <NA>   <NA>

Upvotes: 0

Sotos
Sotos

Reputation: 51582

An idea is to split the Class by every character and then use Reduce with accumulate = TRUE in order to paste them back together one-by-one. We then set their length to the maximum length, rbind and cbind back to the original data frame, i.e.

l1 <- lapply(strsplit(as.character(df$Class), ''), function(i){i1 <- Reduce(paste0, i, accumulate = TRUE); 
                                                               i1 <- i1[!grepl('\\.$', i1)]; 
                                                               i1})
final_list <- lapply(l1, `length<-`, max(lengths(l1)))
cbind.data.frame(df$Site, do.call(rbind, final_list))

which gives,

  df$Site 1    2    3     4
1      A1 D   D2 D2.1 D2.13
2      A2 E   E1 E1.4  <NA>
3      A3 F   FA FA.1  <NA>
4      A4 H   H2 H2.1 H2.14
5      A5 F <NA> <NA>  <NA>
6      AR G   G1 <NA>  <NA>

You can tidy your column names as per usual

Upvotes: 2

boski
boski

Reputation: 2467

Easy with dplyr

df%>%rowwise()%>%mutate(Class1=substr(Class,1,1),
                        Class2=ifelse(nchar(strsplit(Class,"\\.")[[1]][1])==2,substr(Class,1,2),NA),
                        Class3=ifelse(nchar(strsplit(Class,"\\.")[[1]][2])>0,substr(Class,1,4),NA),
                        Class4=ifelse(nchar(Class)>4,Class,NA)
                        )

Source: local data frame [6 x 6]
Groups: <by row>

# A tibble: 6 x 6
  Site  Class Class1 Class2 Class3 Class4
  <chr> <chr> <chr>  <chr>  <chr>  <chr> 
1 A1    D2.13 D      D2     D2.1   D2.13 
2 A2    E1.4  E      E1     E1.4   NA    
3 A3    FA.1  F      FA     FA.1   NA    
4 A4    H2.14 H      H2     H2.1   H2.14 
5 A5    F     F      NA     NA     NA    
6 AR    G1    G      G1     NA     NA

Upvotes: 0

Related Questions