增加 ifelse True 计数并写入列答案

【问题标题】：Increment the ifelse True count and write to a column增加 ifelse True 计数并写入列
【发布时间】：2020-02-03 23:16:10
【问题描述】：

我有一个小标题，我想在每次满足 if 条件时增加计数并写入列。截至目前，我只能改变字符串。我该怎么做？

treeTable_w_bucket %>%
 mutate(bucket = ifelse(Feature == "g201a", "test", bucket))

Tree Node Feature Bucket
 0   0    g201a   NA
 1    1   re20s   NA
 2    0   g201a   NA 
 3    0   g201a   NA

结果

Tree Node Feature Bucket
 0   0    g201a   bucket_1
 1   0    re20s    NA
 2   0    g201s   bucket_2
 3   0    g201s   bucket_3

【问题讨论】：

标签： r dplyr tibble

【解决方案1】：

我们可以使用replace

library(dplyr)
library(stringr)
treeTable_w_bucket <- treeTable_w_bucket %>% 
        mutate(bucket = replace(Bucket, Feature == 'g201a', 
               str_c('bucket_', seq_len(sum(Feature == 'g201a')))))

或使用base R

i1 <- treeTable_w_bucket$Feature == 'g201a'
treeTable_w_bucket$bucket[i1] <- paste0("bucket_", cumsum(i1)[i1])
treeTable_w_bucket
#  Tree Node Feature Bucket   bucket
#1    0    0   g201a     NA bucket_1
#2    1    1   re20s     NA     <NA>
#3    2    0   g201a     NA bucket_2
#4    3    0   g201a     NA bucket_3

数据

treeTable_w_bucket <- structure(list(Tree = 0:3, Node = c(0L, 1L, 0L, 0L), Feature = c("g201a", 
"re20s", "g201a", "g201a"), Bucket = c(NA, NA, NA, NA)), class = "data.frame", row.names = c(NA, 
-4L))

【讨论】：

【解决方案2】：

使用cumsum我们可以计算Feature = "g201a"和replace不是"g201a"到NA的值的次数。

df$Bucket <- with(df, replace(paste0('bucket_', cumsum(Feature == "g201a")), 
                       Feature != "g201a", NA))

#  Tree Node Feature   Bucket
#1    0    0   g201a bucket_1
#2    1    1   re20s     <NA>
#3    2    0   g201a bucket_2
#4    3    0   g201a bucket_3

这也可以在dplyr 管道中完成

library(dplyr)
df %>%
  mutate(Bucket = replace(paste0('bucket_', cumsum(Feature == "g201a")), 
                          Feature != "g201a", NA))

数据

df <- structure(list(Tree = 0:3, Node = c(0L, 1L, 0L, 0L), Feature = structure(c(1L,
2L, 1L, 1L), .Label = c("g201a", "re20s"), class = "factor")), row.names = c(NA,
-4L), class = "data.frame")

【讨论】：