[R] Search for common character strings within a column

[arun]

HI,
In case if you wanted to take "BC" and "CB" as the same.
dat1<- read.table(text="
   Seq,Output
 A B B C D A C,Yes
 B C A C B D A C,Yes
C D A A C D,No
",sep=",",header=TRUE,stringsAsFactors=FALSE)
lapply(str_split(str_trim(dat1$Seq)," ")[dat1$Output=="Yes"],function(x) {x1<-t(combn(x,2)); x2<-sapply(strsplit(apply(x1,1,paste0,collapse=""),""),function(x) paste(x[order(x)],collapse="")); table(x2)}) 
[[1]]
#x2
#AA AB AC AD BB BC BD CC CD 
# 1  4  4  2  1  4  2  1  2 

#[[2]]
#x2
#AA AB AC AD BB BC BD CC CD 
# 1  4  6  2  1  6  2  3  3 
dat1$MaxCombn<- NA
res1<-sapply(str_split(str_trim(dat1$Seq)," ")[dat1$Output=="Yes"],function(x) {x1<-t(combn(x,2)); x2<-sapply(strsplit(apply(x1,1,paste0,collapse=""),""),function(x) paste(x[order(x)],collapse="")); x3<-table(x2); x3[x3%in% max(x3)]}) 

dat1$MaxCombn[dat1$Output=="Yes"]<-lapply(res1,names)
 dat1
#               Seq Output   MaxCombn
#1    A B B C D A C    Yes AB, AC, BC
#2  B C A C B D A C    Yes     AC, BC
#3      C D A A C D     No         NA

A.K.




----- Original Message -----
From: arun <smartpink111 at yahoo.com>
To: R help <r-help at r-project.org>
Cc: 
Sent: Friday, April 12, 2013 4:37 PM
Subject: Re: Search for common character strings within a column 

Hi,
May be this helps:
Not sure how you wanted to select those two letters.  


dat1<- read.table(text="
   Seq,Output
 A B B C D A C,Yes
 B C A C B D A C,Yes
C D A A C D,No
",sep=",",header=TRUE,stringsAsFactors=FALSE)
library(stringr)
lapply(str_split(str_trim(dat1$Seq)," ")[dat1$Output=="Yes"],function(x) {x1<-t(combn(x,2)); apply(x1,1,paste0,collapse="")})
#[[1]]
# [1] "AB" "AB" "AC" "AD" "AA" "AC" "BB" "BC" "BD" "BA" "BC" "BC" "BD" "BA" "BC"
#[16] "CD" "CA" "CC" "DA" "DC" "AC"

#[[2]]
# [1] "BC" "BA" "BC" "BB" "BD" "BA" "BC" "CA" "CC" "CB" "CD" "CA" "CC" "AC" "AB"
#[16] "AD" "AA" "AC" "CB" "CD" "CA" "CC" "BD" "BA" "BC" "DA" "DC" "AC"

res<- sapply(str_split(str_trim(dat1$Seq)," ")[dat1$Output=="Yes"],function(x) {x1<-t(combn(x,2)); x2<-table(apply(x1,1,paste0,collapse="")); x2[which.max(x2)]})
res
#BC BC 
# 4  4
 
dat1$MaxCombn<-NA
 dat1$MaxCombn[dat1$Output=="Yes"]<- names(res)
 dat1
#               Seq Output MaxCombn
#1    A B B C D A C    Yes       BC
#2  B C A C B D A C    Yes       BC
#3      C D A A C D     No     <NA>
A.K.


>I have a dataset (data) that consists of two columns: Seq and output. 
Each entry in Seq is a combination of As,Bs,Cs and Ds and ranges from 5 – >30 characters in length. Each sequence is associated with an output of 
either yes or no such that: 
>
 >     Seq                  Output 
>(1) A B B C D A C             Yes 
>(2) B C A C B D A C    Yes 
>(3) C D A A C D         No 
>
>etc, etc. 
>
>I want to find which 2 letter (A B, A C, A D, etc) strings are 
most associated with each output. Essentially I want to find which 2 
letter combinations >occur most frequently in the column Seq, when the 
output is Yes. I’m new to R and can’t figure out a solution to this 
problem. 
>
>Any help greatly appreciated! 
>
>Cheers, 
>
>AB