尝试在数据框 R 组内按顺序标记

Question

我有我的数据框的一个子集：

df = data.frame(retailer_id = c(1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2),
                store_id = c(166, 166, 167, 167, 167, 167, 167, 167, 167, 167, 167, 167, 167, 167, 167, 167, 167, 167, 167, 167),
                quad_id = c(2017010104, 2017012904, 2017010104, 2017012904, 2017022604, 2017032604 ,2017042304, 2017052104, 2017061804, 
                            2017071604, 2017081304, 2017091004, 2017100804, 2017110504, 2017120304, 2017123104, 2018012804, 2018022504, 2018032504, 2018042204))

其中 2017010104 对应日期 01/01/2017 等等。我试图参考年份按顺序标记这些不同的 quad_ids。因此，例如我正在尝试获取输出：

df = data.frame(retailer_id = c(1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2),
                store_id = c(166, 166, 167, 167, 167, 167, 167, 167, 167, 167, 167, 167, 167, 167, 167, 167, 167, 167, 167, 167),
                quad_id = c(2017010104, 2017012904, 2017010104, 2017012904, 2017022604, 2017032604 ,2017042304, 2017052104, 2017061804, 
                            2017071604, 2017081304, 2017091004, 2017100804, 2017110504, 2017120304, 2017123104, 2018012804, 2018022504, 2018032504, 2018042204),
                Snum = c(1,  2,  1,  2,  3,  4,  5,  6,  7,  8,  9, 10, 11, 12, 13, 14, 1, 2, 3, 4))

您可以看到 retailer_id = 2，store_id = 167，2017 年的周标记为 1-14，然后当周从 2018 年开始时，它开始按顺序从1 直到达到该分组中从 2019 年开始的一周。

我试过了：

DT <- data.table(df)

DT[, Snum := seq_len(.N), by = list(retailer_id, store_id)]

然而，这不是按年份顺序标记，而是按store_id顺序标记。有没有办法来解决这个问题？（此示例代码仅显示两个零售商和两个商店，而我的实际数据框和数百个不同的零售商和商店）

Answer 1

这是一个使用 tidyverse

的解决方案

df = data.frame(retailer_id = c(1, 1, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, 2),
                store_id = c(166, 166, 167, 167, 167, 167, 167, 167, 167, 167, 167, 167, 167, 167, 167, 167, 167, 167, 167, 167),
                quad_id = c(2017010104, 2017012904, 2017010104, 2017012904, 2017022604, 2017032604 ,2017042304, 2017052104, 2017061804, 
                            2017071604, 2017081304, 2017091004, 2017100804, 2017110504, 2017120304, 2017123104, 2018012804, 2018022504, 2018032504, 2018042204))

library(tidyverse)

getYear = function(x) {
  x %>% 
    str_extract("^\d{4}") %>% 
    as.integer() %>% 
    return()
}

tmp = df %>% 
  mutate(year = getYear(quad_id)) %>% 
  group_by(year, retailer_id, store_id) %>% 
  mutate(Snum = 1:n())

> tmp
# A tibble: 20 x 5
# Groups:   year, retailer_id, store_id [3]
   retailer_id store_id    quad_id  year  Snum
         <dbl>    <dbl>      <dbl> <int> <int>
 1           1      166 2017010104  2017     1
 2           1      166 2017012904  2017     2
 3           2      167 2017010104  2017     1
 4           2      167 2017012904  2017     2
 5           2      167 2017022604  2017     3
 6           2      167 2017032604  2017     4
 7           2      167 2017042304  2017     5
 8           2      167 2017052104  2017     6
 9           2      167 2017061804  2017     7
10           2      167 2017071604  2017     8
11           2      167 2017081304  2017     9
12           2      167 2017091004  2017    10
13           2      167 2017100804  2017    11
14           2      167 2017110504  2017    12
15           2      167 2017120304  2017    13
16           2      167 2017123104  2017    14
17           2      167 2018012804  2018     1
18           2      167 2018022504  2018     2
19           2      167 2018032504  2018     3
20           2      167 2018042204  2018     4

请注意，如果您的数据未按 retailer_id、store_id 和 year 排序，则会导致问题。

Answer 2

我们可以使用 stringr 包中的 str_match 和正则表达式 '^[[:digit:]]{4}' 来匹配前四位数字：

library(dplyr)
library(stringr)
df %>% 
    group_by(Snum = str_match(quad_id, '^[[:digit:]]{4}')) %>% 
    mutate(Snum = row_number())

输出：

retailer_id store_id    quad_id  Snum
         <dbl>    <dbl>      <dbl> <int>
 1           1      166 2017010104     1
 2           1      166 2017012904     2
 3           2      167 2017010104     3
 4           2      167 2017012904     4
 5           2      167 2017022604     5
 6           2      167 2017032604     6
 7           2      167 2017042304     7
 8           2      167 2017052104     8
 9           2      167 2017061804     9
10           2      167 2017071604    10
11           2      167 2017081304    11
12           2      167 2017091004    12
13           2      167 2017100804    13
14           2      167 2017110504    14
15           2      167 2017120304    15
16           2      167 2017123104    16
17           2      167 2018012804     1
18           2      167 2018022504     2
19           2      167 2018032504     3
20           2      167 2018042204     4

尝试在数据框 R 组内按顺序标记

Trying to label sequentially within groups of dataframe R

r

dataframe

label

grouping

sequential