How to split a character vector into data frame?

问题:

I'm still relatively new to R and hope you can again help me. I have a character vector with a length of 42000. The vector looks like this:

a <- c("blablabla-19960101T000000Z-1.tsv", "blablabla-19960101T000000Z-2.tsv", "blablabla-19960101T000000Z-3.tsv")

I want to split the vector into a data frame which looks like this:

Name        Date          no
blablabla   1996-01-01    1
blablabla   1996-01-01    2
blablabla   1996-01-01    3

I'm struggling with the splitting as well as the creation of my data frame. Can someone help me with this? Thanks!

回答1:

DF <- data.frame(do.call(rbind, strsplit(a, "-", fixed=TRUE)))
DF[,2] <- as.Date(DF[,2] , format="%Y%m%d")
DF[,3] <- as.integer(gsub(".tsv", "", DF[,3], fixed=TRUE))

#         X1         X2 X3
#1 blablabla 1996-01-01  1
#2 blablabla 1996-01-01  2
#3 blablabla 1996-01-01  3

回答2:

maybe with

library(reshape2)
colsplit(a, "\\-", names=c("A", "B", "C"))

          A                B     C
1 blablabla 19960101T000000Z 1.tsv
2 blablabla 19960101T000000Z 2.tsv
3 blablabla 19960101T000000Z 3.tsv

or

b <- colsplit(a, "[[:punct:]]|\\T|\\.", names=c("A", "B", "C", "D","E"))

          A        B       C D   E
1 blablabla 19960101 000000Z 1 tsv
2 blablabla 19960101 000000Z 2 tsv
3 blablabla 19960101 000000Z 3 tsv

and then

library(lubridate)
b$B <- ymd(b$B)

          A          B       C D   E
1 blablabla 1996-01-01 000000Z 1 tsv
2 blablabla 1996-01-01 000000Z 2 tsv
3 blablabla 1996-01-01 000000Z 3 tsv

str(b)
'data.frame':   3 obs. of  5 variables:
 $ A: chr  "blablabla" "blablabla" "blablabla"
 $ B: POSIXct, format: "1996-01-01" "1996-01-01" "1996-01-01"
 $ C: chr  "000000Z" "000000Z" "000000Z"
 $ D: int  1 2 3
 $ E: chr  "tsv" "tsv" "tsv"

回答3:

You can almost use read.table directly, but your date format isn't the same as what R would use for the colClasses argument.

No problem. Just specify your own class and proceed :-)

## Create a class called "ymdDate"
setClass("ymdDate")
setAs("character", "ymdDate", function(from) as.Date(from, format="%Y%m%d"))

## Use `read.table` on your character vector. For convenience, I've
##   used `gsub` to get rid of the `.tsv` in before reading it in.
out <- read.table(text = gsub(".tsv$", "", a), header = FALSE, 
                  sep = "-", colClasses=c("character", "ymdDate", "integer"))
out
#          V1         V2 V3
# 1 blablabla 1996-01-01  1
# 2 blablabla 1996-01-01  2
# 3 blablabla 1996-01-01  3
str(out)
# 'data.frame':  3 obs. of  3 variables:
#  $ V1: chr  "blablabla" "blablabla" "blablabla"
#  $ V2: Date, format: "1996-01-01" "1996-01-01" "1996-01-01"
#  $ V3: int  1 2 3

回答4:

I know I'm late to this party, but I wanted to see this same idea in a magrittr pipe and using more tidyverse functions. Here's what I've got:

library(stringr)
library(lubridate)
library(tidyverse)

a <- c("blablabla-19960101T000000Z-1.tsv", "blablabla-19960101T000000Z-2.tsv", "blablabla-19960101T000000Z-3.tsv")

a %>%
strsplit('-') %>%
transpose() %>%
map_dfc(~data_frame(.x)) %>%
unnest() %>%
set_names(c('Name','Date','no')) %>% 
mutate(Date = Date %>%
        str_extract('\\d+') %>% 
        ymd(),
        no = str_extract(no, '\\d+'))