Question

我想找到某个数据帧中出现次数最多的数据帧的列。例如，以下数据框：

as.data.frame(cbind(c(1,4,6,9,20),c(2,4,7,7,3),c(4,7,6,4,2),c(1,4,6,9,20),c(4,7,6,4,2),c(7,4,6,4,2)))

我的第一个想法是对数据框中的列进行排序，这样我得到： as.data.frame(cbind(c(1,4,6,9,20),c(2,3,4,7,7),c(2,4,4,6,7),c(1,4,6,9,20),c(2,4,4,6,7),c(2,4,4,6,7))) 然后在此数据框中找到出现次数最多的列。然后它将返回c(2,4,4,6,7)。如何在R中完成此操作？

Answer 1

基本上，您可以paste并将数字计数在table中，然后选择which.max。

d1[, which.max(table(sapply(d1, paste, collapse="")))]
# [1] 2 4 4 6 7

数据

d1 <- structure(list(X1 = c(1, 4, 6, 9, 20), X2 = c(2, 3, 4, 7, 7), 
X3 = c(2, 4, 4, 6, 7), X4 = c(1, 4, 6, 9, 20), X5 = c(2, 
4, 4, 6, 7), X6 = c(2, 4, 4, 6, 7)), class = "data.frame", row.names = c(NA, -5L))

Answer 2

与@ jay.sf基本相同的解决方案，但使用tidyverse：

library(dplyr)
#> 
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#> 
#>     filter, lag
#> The following objects are masked from 'package:base':
#> 
#>     intersect, setdiff, setequal, union

dataset <- as.data.frame(x = cbind(c(1, 4, 6, 9, 20),
                                   c(2, 4, 7, 7, 3),
                                   c(4, 7, 6, 4, 2),
                                   c(1, 4, 6, 9, 20),
                                   c(4, 7, 6, 4, 2),
                                   c(7, 4, 6, 4, 2)))

dataset %>%
  summarise_all(.funs = ~ paste0(sort(.), collapse = "")) %>%
  as.numeric() %>%
  table() %>%
  which.max() %>%
  `[`(dataset, .)
#>   V2
#> 1  2
#> 2  4
#> 3  7
#> 4  7
#> 5  3

^{由reprex package（v0.3.0）于2019-06-15创建}

Answer 3

如果要匹配多个列：

# Creating a table of pasted & sorted column values
counts_df1 <- table(do.call(paste, data.frame(t(sapply(df1, sort)))))

# If you want the sorted order returned as a single element vector:
names(counts_df1[counts_df1 == max(counts_df1)])
[1] "2 4 4 6 7"

或者，您可以这样做来索引数据框中的列：

# Creating collapsed strings from columns
df1_vec <- sapply(df1, function(x) paste0(sort(x), collapse = ""))

# Counting the frequency of each collapsed strings
df1_colsum <- colSums(outer(df1_vec, df1_vec, `==`))

# Subsetting the dataframe based on the most frequent columns that are not duplicates
df1[, df1_colsum == max(df1_colsum) & !duplicated(df1_vec)]
[1] 4 7 6 4 2

查找数据框中出现次数最多的列

3 个答案: