R中按列汇总数据答案

【问题标题】：Summary data by column in RR中按列汇总数据
【发布时间】：2021-04-23 00:01:34
【问题描述】：

我有以下数据

pt_id <- c(1,1,1,1,1,2,2,2,3,3,3,3,3,4,4,4,4)
Tob_pk <- c(2, 5, 7, 1, 8, 12, 14, 3, 6, 8, 10, 20, 13, 5, 4, 12, 10)
Tobacco <- c("Once","Twice","Never", NA, NA, NA, NA, NA,"Once","Twice","Quit","Once",NA,NA,"Never", NA, "Never")
Alcohol <- c("Twice", "Once",NA, NA, "Never", NA, NA, "Once", NA, "Quit", "Twice", NA, "Once", NA, NA, "Never", "Never")
PA <- c("Once",NA,"Never", NA, NA, NA, NA, NA,"Once",NA,"Quit","Once",NA,NA,"Never", NA, NA)
mydata <- data.frame(pt_id, Tob_pk, Tobacco, Alcohol, PA)
mydata

我想要数据集中每个变量的摘要/比例，我尝试使用以下代码获取每个变量的摘要/比例

data_summ <- mydata %>%
  summarize_at(.vars=3:5, funs(prop.table(.)))

但是，我收到以下错误

Error: Problem with `summarise()` input `Tobacco`.
x invalid 'type' (character) of argument
ℹ Input `Tobacco` is `prop.table(Tobacco)`.
Run `rlang::last_error()` to see where the error occurred.

我不确定我哪里出错了。如果我能得到任何建议来获得以下输出但也有 NA 的百分比，那将很有帮助。

Tobacco                   Alcohol                        PA

Never    0.3333333      Never    0.3333333       Never    0.3333333  
Once     0.3333333      Once     0.3333333       Once     0.5000000
Quit     0.1111111      Quit     0.1111111       Quit.    0.1666667
Twice    0.2222222      Twice    0.2222222

提前致谢！

【问题讨论】：

标签： r summarize

【解决方案1】：

使用base

pt_id <- c(1,1,1,1,1,2,2,2,3,3,3,3,3,4,4,4,4)
Tob_pk <- c(2, 5, 7, 1, 8, 12, 14, 3, 6, 8, 10, 20, 13, 5, 4, 12, 10)
Tobacco <- c("Once","Twice","Never", NA, NA, NA, NA, NA,"Once","Twice","Quit","Once",NA,NA,"Never", NA, "Never")
Alcohol <- c("Twice", "Once",NA, NA, "Never", NA, NA, "Once", NA, "Quit", "Twice", NA, "Once", NA, NA, "Never", "Never")
PA <- c("Once",NA,"Never", NA, NA, NA, NA, NA,"Once",NA,"Quit","Once",NA,NA,"Never", NA, NA)
mydata <- data.frame(pt_id, Tob_pk, Tobacco, Alcohol, PA)

apply(mydata[3:5], 2, function(x) prop.table(table(x, useNA="ifany")))
$Tobacco
x
     Never       Once       Quit      Twice       <NA> 
0.17647059 0.17647059 0.05882353 0.11764706 0.47058824 

$Alcohol
x
     Never       Once       Quit      Twice       <NA> 
0.17647059 0.17647059 0.05882353 0.11764706 0.47058824 

$PA
x
     Never       Once       Quit       <NA> 
0.11764706 0.17647059 0.05882353 0.64705882

^{由reprex package (v0.3.0) 于 2021-01-18 创建}

使用tidyverse

library(tidyverse)
map_dfr(mydata[3:5], ~prop.table(table(.x))) 
#> # A tibble: 3 x 4
#>   Never     Once      Quit      Twice    
#>   <table>   <table>   <table>   <table>  
#> 1 0.3333333 0.3333333 0.1111111 0.2222222
#> 2 0.3333333 0.3333333 0.1111111 0.2222222
#> 3 0.3333333 0.5000000 0.1666667        NA

^{由reprex package (v0.3.0) 于 2021-01-18 创建}

【讨论】：

感谢您的帮助！在我的原始数据集中，我有近 35 个变量/列，其中一些是“dbl”变量而不是字符变量。有没有一种方法可以在我的代码中跳过那些带有“dbl”变量的列，而不必手动选择一个范围（例如，3:5 然后 8:5）？
我们可以将基础 R 代码的输出导出到 excel 文件中吗？
试试这个map_dfr(.x = select(mydata, where(is.character)), .f = prop.table(table(.x)))
for base cols_char <- sapply(mydata, is.character) apply(mydata[, cols_char], 2, function(x) prop.table(table(x, useNA="ifany")))

【解决方案2】：

使用 dplyr：

library(tidyverse)

df <- mydata %>%
  select(3:5) %>%
  gather('Your_vice', 'freq', 1:3) %>%
  group_by(Your_vice, freq) %>%
  summarize(n = n()) %>%
  mutate(perc = n/sum(n))

df

# A tibble: 14 x 4
# Groups:   Your_vice [3]
   Your_vice freq      n   perc
   <chr>     <chr> <int>  <dbl>
 1 Alcohol   Never     3 0.176 
 2 Alcohol   Once      3 0.176 
 3 Alcohol   Quit      1 0.0588
 4 Alcohol   Twice     2 0.118 
 5 Alcohol   NA        8 0.471 
 6 PA        Never     2 0.118 
 7 PA        Once      3 0.176 
 8 PA        Quit      1 0.0588
 9 PA        NA       11 0.647 
10 Tobacco   Never     3 0.176 
11 Tobacco   Once      3 0.176 
12 Tobacco   Quit      1 0.0588
13 Tobacco   Twice     2 0.118 
14 Tobacco   NA        8 0.471

【讨论】：

【解决方案3】：

您可以将变量转换为具有相应水平的因子。然后table 和useNA="ifany" 选项显示NA。

mydata[3:5] <- lapply(mydata[3:5], factor, levels=c("Never", "Once", "Quit", "Twice"))
res1 <- sapply(mydata[3:5], function(x) prop.table(table(x)))
res1
#         Tobacco   Alcohol        PA
# Never 0.3333333 0.3333333 0.3333333
# Once  0.3333333 0.3333333 0.5000000
# Quit  0.1111111 0.1111111 0.1666667
# Twice 0.2222222 0.2222222 0.0000000

res2 <- sapply(mydata[3:5], function(x) prop.table(table(x, useNA="ifany")))
res2
#          Tobacco    Alcohol         PA
# Never 0.17647059 0.17647059 0.11764706
# Once  0.17647059 0.17647059 0.17647059
# Quit  0.05882353 0.05882353 0.05882353
# Twice 0.11764706 0.11764706 0.00000000
# <NA>  0.47058824 0.47058824 0.64705882

【讨论】：

输出可以导出到excel文件吗？我尝试使用write.xlsx(res2, file = "myDATA.xlsx", sheetName = "res2", col.names = TRUE, row.names = TRUE, append = FALSE)，但它不起作用！

【解决方案4】：

这是使用prop.table + table + na.omit + factor 的基本 R 选项

do.call(
  cbind,
  lapply(
    mydata[3:5],
    function(x) {
      prop.table(
        table(
          na.omit(
            factor(x, levels = unique(na.omit(unlist(mydata[3:5]))))
          )
        )
      )
    }
  )
)

给了

        Tobacco   Alcohol        PA
Once  0.3333333 0.3333333 0.5000000
Twice 0.2222222 0.2222222 0.0000000
Never 0.3333333 0.3333333 0.3333333
Quit  0.1111111 0.1111111 0.1666667

【讨论】：