使用gsub和regexpr在r中子字符串化

Question

I need to change the following

test <- c("August 08, 2016, Hour 23",
          "June 26, 2016, Hour 14",
          "November 26, 2016, Hour 01")

test1 <- c("Wednesday:8pm-12pm:31days",
"Tuesday:7pm-10pm:6days|Today:7AM-6PM:7days")

Edit:- In test1, I don't really care much about the day of the week, but am more interested in the timestamp. I would like to see 8PM-12PM converted into 24 hr time format as : 2000 - am agreeable with a string as an output as I require a 4 digit number. (Anything before 10 AM would need to be 0x)

into two datasets as:-

a$date <- c(08/08/2016,06/26/2016,11/26/2016) # all in date class
a$hour <- c(23, 14 , 01) #all should be numeric


b$time <- c("2000","1922","0718") #can be character
b$days <- c(31,6,7)  #needs to be numeric

The logic for the hour and days cases would be similar. I'm looking to use gsub and regexpr in R.

My current process for the date section is too long and tedious:-

mat <- as.data.frame(matrix(unlist(strsplit(test," ")),ncol=5,byrow=T))

mat$V6 <-  str_replace_all(paste(as.numeric(str_replace_all(mat$V2,"[[:punct:]]","")),
                          "-",as.character(mat$V1),
                          "-",as.numeric(str_replace_all(mat$V3,"[[:punct:]]",""))),
                          "[[:space:]]","")


mat$V7 <- as.Date(mat$V6, format="%d-%B-%Y")

class(mat$V7)

mat$V8 <- as.numeric(as.character(mat$V5))

Any suggestions for using gsub and regexpr in both cases would be appreciated.

Answer 1

This does the same thing as your mat line. Go ahead and try it.

library(reshape2)
mat <- colsplit(test," ", c("M","D","YYYY","HR","Time"))

I think is your best bet, instead of using gsub or regexpr.

mat$Len <- paste(mat$D,mat$M,mat$YYYY)
mat$Len <- gsub(",","",gsub(" ","-",mat$Len))

I am not a fan of using nested gsub's but it serves a purpose here. Keeps this a bit more concise. This should take care of the mat$v6 line.