基于r中两列的频率计数

Question

我只有一个数据框如下。

df=data.frame(o=c(rep("a",12),rep("b",3)), d=c(0,0,1,0,0.3,0.6,0,1,2,3,4,0,0,1,0))

> df
   o   d
1  a 0.0
2  a 0.0
3  a 1.0
4  a 0.0
5  a 0.3
6  a 0.6
7  a 0.0
8  a 1.0
9  a 2.0
10 a 3.0
11 a 4.0
12 a 0.0
13 b 0.0
14 b 1.0
15 b 0.0

我想添加一个新列来计算基于 'o' 和 'd' 两列的频率。如果列 'd' 的值为零，则频率应重新从 1 开始，如下所示（手工制作）。

> df_result
   o   d freq
1  a 0.0    1
2  a 0.0    2
3  a 1.0    2
4  a 0.0    3
5  a 0.3    3
6  a 0.6    3
7  a 0.0    5
8  a 1.0    5
9  a 2.0    5
10 a 3.0    5
11 a 4.0    5
12 a 0.0    1
13 b 0.0    2
14 b 1.0    2
15 b 0.0    1

Answer 1

在基础 R 中，使用 ave :

df$freq <- with(df, ave(d, cumsum(d == 0), FUN = length))
df

#   o   d freq
#1  a 0.0    1
#2  a 0.0    2
#3  a 1.0    2
#4  a 0.0    3
#5  a 0.3    3
#6  a 0.6    3
#7  a 0.0    5
#8  a 1.0    5
#9  a 2.0    5
#10 a 3.0    5
#11 a 4.0    5
#12 a 0.0    1
#13 b 0.0    2
#14 b 1.0    2
#15 b 0.0    1

与dplyr :

library(dplyr)
df %>% add_count(grp = cumsum(d == 0))

Answer 2

使用data.tables和@Ronak Shah方法

df=data.frame(o=c(rep("a",12),rep("b",3)), d=c(0,0,1,0,0.3,0.6,0,1,2,3,4,0,0,1,0)) 

library(data.table)

setDT(df)[, freq := .N, by = cumsum(d == 0)]

df
#>     o   d freq
#>  1: a 0.0    1
#>  2: a 0.0    2
#>  3: a 1.0    2
#>  4: a 0.0    3
#>  5: a 0.3    3
#>  6: a 0.6    3
#>  7: a 0.0    5
#>  8: a 1.0    5
#>  9: a 2.0    5
#> 10: a 3.0    5
#> 11: a 4.0    5
#> 12: a 0.0    1
#> 13: b 0.0    2
#> 14: b 1.0    2
#> 15: b 0.0    1

^{由 reprex package (v1.0.0)}

创建于 2021-02-26

Answer 3

使用 rle()

的另一个答案

df$freq <- with(rle(cumsum(df$d == 0)), rep(lengths, lengths))
df

   o   d freq
1  a 0.0    1
2  a 0.0    2
3  a 1.0    2
4  a 0.0    3
5  a 0.3    3
6  a 0.6    3
7  a 0.0    5
8  a 1.0    5
9  a 2.0    5
10 a 3.0    5
11 a 4.0    5
12 a 0.0    1
13 b 0.0    2
14 b 1.0    2
15 b 0.0    1

基于r中两列的频率计数

Frequency count based on two columns in r

r

frequency

count

dataframe