Question

如何使用dplyr在R中执行此操作。我有这张桌子

Product   Week    Count
A         Wk1     2 
A         Wk2     2 
B         Wk1     3
C         Wk2     4
C         Wk3     3

我希望输出为此

Product   Week   Count
A         Wk1     2
A         Wk2     2
A         Wk3     0
B         Wk1     3
B         Wk2     0
B         Wk3     0
C         Wk1     0
C         Wk2     4
C         Wk3     3

Answer 1

使用tidyverse包，以下代码将完成您缺少的组合，允许您指定缺少组合的值：

library(tidyverse)
my_df %>%
    complete(Product,Week,fill=list(Count=0))

结果如下：

# A tibble: 9 x 3
  Product Week  Count
  <chr>   <chr> <dbl>
1 A       Wk1    2.00
2 A       Wk2    2.00
3 A       Wk3    0   
4 B       Wk1    3.00
5 B       Wk2    0   
6 B       Wk3    0   
7 C       Wk1    0   
8 C       Wk2    4.00
9 C       Wk3    3.00

Answer 2

这是为缺少产品周组合添加零计数行的一种方法：

# 0. attach dplyr for join, mutate, and magrittr pipe (%>%)
library(dplyr)

# 1. get the data as a data frame 
dat <- read.table(text="
Product   Week    Count
A         Wk1     2 
A         Wk2     2 
B         Wk1     3
C         Wk2     4
C         Wk3     3
", header=TRUE)

# 2. get unique products and unique weeks 
products <- unique(dat$Product)
weeks <- unique(dat$Week)

# 3.1 create a df with all combos of product and week 
expand.grid(Product=products, Week=weeks, stringsAsFactors=FALSE) %>%
  # 3.2 join it with original data (introduces NA counts for missing combos)
  left_join(dat, by=c("Product", "Week")) %>% 
  # 3.3 replace NA counts with zero 
  mutate(Count = ifelse(is.na(Count), 0, Count))

确实有其他方法可以做到这一点，但是这样做（只要您在原始表格中永远不会有NA个值。）

完成该组然后将零添加到丢失的单元格

2 个答案: