R 统计ABC的XXXX后查询日期的所有条目必须按其用途、分组比（）划分，并详细说明数据

我打算列出每个ID中查询ABC后的所有用途（其中XXXX表示其他公司）。 示例表如下所示：

     ID            Company  INQUIRY-DATE Purpose
    A15217177635833 XXXX    25-08-2018   X
    A15217177635833 ABC     28-06-2018   Y
    A15217177635833 XXXX    05-05-2018   Z
    A15217177635833 XXXX    28-05-2019   A
    F15039820795577 ABC     22-08-2017   X
    F15039820795577 XXXX    15-06-2017   Y
    F15039820795577 XXXX    15-08-2018   Z
    F15039820795577 XXXX    25-08-2018   Z
    F15039820795577 XXXX    15-08-2018   A

预期产出：

     ID             Count_Z  Count_A
    A15217177635833 1         1
    F15039820795577 2         1

这意味着ABC的XXXX查询后日期的所有条目必须根据其用途进行划分。 我尝试了使用group by和mutate（count_z），但没有成功

---

我不知道如何在分组依据后获得详细数据，因为我的知识分组依据与摘要一起使用。

我们首先将

查询日期

转换为日期对象，

按
ID
排列

数据，

查询日期

对于每个分组，只选择第一次出现

“ABC”

之后的行，对每个

目的进行计数

，然后以广泛的格式传播数据

library(dplyr)

df %>%
   mutate(`INQUIRY-DATE` = as.Date(`INQUIRY-DATE`, "%d-%m-%Y")) %>%
   arrange(ID, `INQUIRY-DATE`) %>%
   group_by(ID) %>%
   filter(Company != "ABC" & row_number() > match("ABC", Company)) %>%
   count(ID, Purpose) %>%
   tidyr::pivot_wider(names_from = Purpose, values_from = n, 
                      values_fill = list(n = 0))

#  ID                  A     X     Z
#  <fct>           <int> <int> <int>
#1 A15217177635833     1     1     0
#2 F15039820795577     1     0     2

库（dplyr）
df%>%
变异（`INQUIRY-DATE`=as.DATE（`INQUIRY-DATE`，%d-%m-%Y））%>%
安排（ID，`INQUIRY-DATE`）%>%
分组依据（ID）%>%
过滤器（公司！=“ABC”&行号（）>匹配（“ABC”，公司））%>%
计数（ID，用途）%>%
tidyr:：pivot\u wide（名称\u from=Purpose，值\u from=n，
值\u填充=列表（n=0））
#ID A X Z
#               
#1 A15217177635833 1 1 0
#2 F15039820795577 1 0 2

数据

df <- structure(list(ID = structure(c(1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L, 
2L), .Label = c("A15217177635833", "F15039820795577"), class = "factor"), 
Company = structure(c(2L, 1L, 2L, 2L, 1L, 2L, 2L, 2L, 2L), .Label = c("ABC", 
"XXXX"), class = "factor"), `INQUIRY-DATE` = structure(c(5L, 
7L, 1L, 6L, 4L, 2L, 3L, 5L, 3L), .Label = c("05-05-2018", 
"15-06-2017", "15-08-2018", "22-08-2017", "25-08-2018", "28-05-2019", 
"28-06-2018"), class = "factor"), Purpose = structure(c(2L, 
3L, 4L, 1L, 2L, 3L, 4L, 4L, 1L), .Label = c("A", "X", "Y", 
"Z"), class = "factor")), class = "data.frame", row.names = c(NA, -9L))

---

df还有另一种方法。它假定行是按时间顺序排列的

library(tidyr)

xy <- read.table(text = "     ID            Company  INQUIRY-DATE Purpose
    A15217177635833 XXXX    25-08-2018   X
    A15217177635833 ABC     28-06-2018   Y
    A15217177635833 XXXX    05-05-2018   Z
    A15217177635833 XXXX    28-05-2019   A
    F15039820795577 ABC     22-08-2017   X
    F15039820795577 XXXX    15-06-2017   Y
    F15039820795577 XXXX    15-08-2018   Z
    F15039820795577 XXXX    25-08-2018   Z
    F15039820795577 XXXX    15-08-2018   A", header = TRUE)

xys <- split(xy, f = xy$ID)
xya <- sapply(xys, FUN = function(x) {
  # This assumes there can be more than one ABC, so start from the first one.
  start <- min(which(x$Company == "ABC"))
  post.abc <- x[(start + 1):nrow(x), ]
  data.frame(ID = unique(x$ID), counts = table(post.abc$Purpose))

}, simplify = FALSE)

out <- do.call(rbind, xya)
rownames(out) <- NULL

spread(out, key = counts.Var1, value = counts.Freq)

               ID A X Y Z
1 A15217177635833 1 0 0 1
2 F15039820795577 1 0 1 2

library（tidyr）
xy