[英]Is there an R function to clean messy salaries in character format?
I have a column of messy salary data.我有一列混乱的工资数据。 I am wondering if there is a package that has a function made specifically for cleaning this type of messy data.我想知道是否有一个 package 有一个 function 专门用于清理这种类型的混乱数据。 My data looks like:我的数据看起来像:
data.frame(salary = c("40,000-60,000", "40-80K", "$100,000",
"$70/hr", "Between $65-80/hour", "$100k",
"50-60,000 a year", "90"))
#> salary
#> 1 40,000-60,000
#> 2 40-80K
#> 3 $100,000
#> 4 $70/hr
#> 5 Between $65-80/hour
#> 6 $100k
#> 7 50-60,000 a year
#> 8 90
Created on 2020-12-16 by the reprex package (v0.3.0)由代表 package (v0.3.0) 于 2020 年 12 月 16 日创建
and I would like the clean column to be a numeric at the annual level.我希望干净的列是年度级别的数字。 I know how to clean this column manually, I'm just wondering if there are any other packages that can help (other than readr::parse_number()
)我知道如何手动清理此列,我只是想知道是否有任何其他软件包可以提供帮助(除了readr::parse_number()
)
The expected output would look like:预期的 output 如下所示:
#> output
#> 1 50000
#> 2 60000
#> 3 100000
#> 4 145600
#> 5 150800
#> 6 100000
#> 7 55000
#> 8 90000
Here are some first steps you can try.以下是您可以尝试的一些初步步骤。 I define two functions: one replaces a k
or K
with three zeros.我定义了两个函数:一个用三个零替换k
或K
The other adds leading zeros if one number is denoted in thousands and the other is not.如果一个数字以千表示而另一个不是,则另一个添加前导零。
rem_k <- function(x) {
sub("(\\d)[kK]", "\\1,000", x)
}
add_zero <- function(x) {
ifelse(grepl("[1-9]0\\-\\d[0,]{2,}", x), sub("([1-9]0)(\\-\\d[0,]{2,})", "\\1,000\\2", x), x)
}
Finally, I remove all non essential characters:最后,我删除了所有非必要字符:
df %>%
mutate(salary2 = gsub("[^0-9,\\-]", "", add_zero(rem_k(salary))))
salary salary2
1 40,000-60,000 40,000-60,000
2 40-80K 40,000-80,000
3 $100,000 100,000
4 $70/hr 70
5 Between $65-80/hour 65-80
6 $100k 100,000
7 50-60,000 a year 50,000-60,000
8 90 90
One option is to create a column 'salary1' with only the digits and the -
, then separate
it to two columns by the -
, mutate
the values of those columns, based on the substring matches in the original column ie K|k
or hr|hour
ie multiply them with 1000 ( K|k
) or for hourly, based on the number of hours for a year, with case_when
and get the rowMeans
of those columns一种选择是创建仅包含数字和-
的列“salary1”,然后通过-
将其separate
为两列,根据原始列中的mutate
匹配,即K|k
或hr|hour
改变这些列的值hr|hour
即将它们乘以 1000 ( K|k
) 或每小时,基于一年的小时数,使用case_when
并获得这些列的rowMeans
library(dplyr)
library(stringr)
library(tidyr)
df1 %>%
mutate(salary1 = str_remove_all(salary, '[^0-9-]+')) %>%
separate(salary1, into = c('salary1', 'salary2'),
convert = TRUE, extra = 'drop') %>%
mutate(across(c(salary1, salary2),
~ case_when(str_detect(salary, "[Kk]") ~ . * 1000,
str_detect(salary, 'hr|hour') ~ . * 40 * 4 * 12,
nchar(.) < 5 ~ as.numeric(str_pad(., pad = '0',
side = 'right', width = 5)),
TRUE ~ as.numeric(.)))) %>%
transmute(output = rowMeans(select(., salary1, salary2), na.rm = TRUE))
-output -输出
# output
#1 50000
#2 60000
#3 100000
#4 134400
#5 139200
#6 100000
#7 55000
#8 90000
声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.