我是R的新手,我有一个包含17列和超过1米行的大数据集。我想通过分隔符'/'将其中一列拆分为4。 R需要永远完成以下命令。有没有更好的方法来完成以下。我尽可能多地提供了代码信息,并希望得到任何帮助。
sample2 <- read.csv("week1.csv", header=TRUE)
summ1 <- subset(sample2,select= -c(3,7), subset =(SPORTS_ID =='1'))
summ1 <- summ1[,-c(1)]
library(splitstackshape)
summ2 <- concat.split.multiple(summ1,2 , "/")
summ2 <- summ2[,-c(1,15)]
summ3 <- concat.split.multiple(summ2,14, "v")
write.csv(summm3, file="test.csv")
答案 0 :(得分:1)
您可以使用strsplit
:
dat <- data.frame(a = c("a/b/c/d",
"e/f/g/h"),
stringsAsFactors = FALSE)
# a
# 1 a/b/c/d
# 2 e/f/g/h
cbind(dat, do.call(rbind, strsplit(dat$a, "/")))
# a 1 2 3 4
# 1 a/b/c/d a b c d
# 2 e/f/g/h e f g h
答案 1 :(得分:1)
正如我在评论中提到的,如果您的数据是平衡的(也就是说,您希望在分割数据后得到一个漂亮的矩形数据集),那么您应该查看我的concat.split.DT
函数。
以下是一些测试。
dat <- do.call(rbind, replicate(1e4, dat, simplify=FALSE))
dim(dat)
# [1] 20000 1
“stringr”函数可能有点慢:
library(stringr)
system.time(do.call(rbind, str_split(dat$a, "/")))
# user system elapsed
# 3.194 0.000 3.211
但其他解决方案如何实现呢?
fun1 <- function() concat.split.multiple(dat, "a", "/")
fun2 <- function() do.call(rbind, strsplit(dat$a, "/", fixed=TRUE))
## ^^ fixed = TRUE will make a big difference
fun3 <- function() concat.split.DT(dat, "a", "/")
library(microbenchmark)
microbenchmark(fun1(), fun2(), fun3(), times = 10)
# Unit: milliseconds
# expr min lq median uq max neval
# fun1() 530.46597 534.13486 535.19139 538.91488 553.61919 10
# fun2() 30.22265 31.07287 31.81474 32.93936 40.28859 10
# fun3() 22.57517 22.94169 23.10297 23.30907 31.97640 10
因此,对于常规concat.split.multiple
(仅在引擎盖下使用read.table
)大约需要半秒钟,而strsplit
和concat.split.DT
的结果要好得多({1}}后者使用了来自“data.table”的fread
。
dat <- do.call(rbind, replicate(50, dat, simplify=FALSE))
dim(dat)
# [1] 1000000 1
microbenchmark(fun2(), fun3(), times = 5)
# Unit: seconds
# expr min lq median uq max neval
# fun2() 6.257892 6.522199 13.728283 13.934860 14.277432 5
# fun3() 1.671739 1.830485 2.203076 2.470872 2.572917 5
concat.split.DT
方法的优点是可以使用简单的语法分割多个列:
dat2 <- do.call(cbind, replicate(5, dat, simplify = FALSE))
dim(dat2)
# [1] 1000000 5
names(dat2) <- make.unique(names(dat2))
head(dat2)
# a a.1 a.2 a.3 a.4
# 1 a/b/c/d a/b/c/d a/b/c/d a/b/c/d a/b/c/d
# 2 e/f/g/h e/f/g/h e/f/g/h e/f/g/h e/f/g/h
# 3 a/b/c/d a/b/c/d a/b/c/d a/b/c/d a/b/c/d
# 4 e/f/g/h e/f/g/h e/f/g/h e/f/g/h e/f/g/h
# 5 a/b/c/d a/b/c/d a/b/c/d a/b/c/d a/b/c/d
# 6 e/f/g/h e/f/g/h e/f/g/h e/f/g/h e/f/g/h
现在,让我们立刻拆分所有这些:
system.time(out <- concat.split.DT(dat2, names(dat2), "/"))
# user system elapsed
# 6.260 0.040 6.532
out
# a_1 a_2 a_3 a_4 a.1_1 a.1_2 a.1_3 a.1_4 a.2_1 a.2_2 a.2_3 a.2_4 a.3_1
# 1: a b c d a b c d a b c d a
# 2: e f g h e f g h e f g h e
# 3: a b c d a b c d a b c d a
# 4: e f g h e f g h e f g h e
# 5: a b c d a b c d a b c d a
# ---
# 999996: e f g h e f g h e f g h e
# 999997: a b c d a b c d a b c d a
# 999998: e f g h e f g h e f g h e
# 999999: a b c d a b c d a b c d a
# 1000000: e f g h e f g h e f g h e
# a.3_2 a.3_3 a.3_4 a.4_1 a.4_2 a.4_3 a.4_4
# 1: b c d a b c d
# 2: f g h e f g h
# 3: b c d a b c d
# 4: f g h e f g h
# 5: b c d a b c d
# ---
# 999996: f g h e f g h
# 999997: b c d a b c d
# 999998: f g h e f g h
# 999999: b c d a b c d
# 1000000: f g h e f g h
答案 2 :(得分:0)
这应该让你开始。您可能需要根据数据包含的内容调整正则表达式模式。可重复的例子会有所帮助。 How to make a great R reproducible example?
library(stringr)
df <- as.data.frame(cbind(x = seq(1,10,1), y = rep("first/second", 10)), stringsAsFactors = FALSE)
df
df$first <- str_replace(df$y, "\\/\\w+", "")
df$second <- str_replace(df$y, "\\w+\\/", "")
df
> df
x y first second
1 1 first/second first second
2 2 first/second first second
3 3 first/second first second
4 4 first/second first second
5 5 first/second first second
6 6 first/second first second
7 7 first/second first second
8 8 first/second first second
9 9 first/second first second
10 10 first/second first second
答案 3 :(得分:0)
如果您打算使用字符并且不介意列表,str_split
包中的stringr
应该有帮助
library(stringr)
x <- 'hello/hi/hey/hola'
str_split(x)
[[1]]
[1] "hello" "hi" "hey" "hola"