Dataframe processing

Question

I have a dataframe, which I read by Match <- read.table(\"Match.txt\", sep=\"\", fill =T, stringsAsFactors = FALSE, quote = \"\", header = F) and looks like

Answer 1

Maybe not the best use of stringr or tidyr, but this can be done in the hadleyverse in a somewhat readable manner...

The logic flow is:

• Determine the group by using tidyr::fill with ifelse("Inspecting", rowname, NA).
• Mutate the fields to what you wanted
• Use reshape (dcast) to get the format that you want.

---

library(dplyr)
library(tidyr)
library(reshape2)
library(stringr)

is_in <- function(v1part) {
  return(ifelse(length(v1part) > 0, "B", "U"))
}

ab1<- ab %>% 
  add_rownames() %>%
  mutate(rowname = ifelse(V1=="Inspecting", rowname, NA),
         V4a = ifelse(V4 == "(-)" | V4 == "(+)", NA, V4),

         chr = str_extract_all(ab$V4, "^chr[^:]+", simplify = T)[,1],
         chr = ifelse(chr=="", NA, chr),

         start = str_split_fixed(V4a, ":|-", 3)[,2],
         start = ifelse(start=="", NA, start), 

         stop = str_split_fixed(V4a, ":|-", 3)[,3],
         stop = ifelse(stop=="", NA, stop),

         V1part = str_split_fixed(V1, "\\$|_", 3)[,2]) %>%
  fill(rowname, .direction="down") %>% 
  group_by(rowname) %>%
  fill(chr, .direction="down") %>%
  fill(start, .direction="down") %>%
  fill(stop, .direction="down") %>%
  dcast(chr+start+stop ~ V1part, fun.aggregate=is_in)

> ab1
   chr     start      stop Var.4 ATF3 CEBPB YY1
1 chr1 173244300 173244500     B    B     B   B
2 chr1 173244350 173244550     B    B     B   U