如何在R中满足特定条件时获得连续出现的平均值

Question

我有一个data ，其中包含按交易日期订购的客户付款时间表。我想计算连续失败付款的平均次数和连续成功付款的平均次数。表格如下所示：

customer_id |transaction_id.|failed_or_success  | transaction_date 
  1         |1              |success            |2021-01-01
  1         |2              |success            |2021-01-15
  1         |3              |failed             |2021-01-30
  1         |4              |success            |2021-02-15

例如，平均连续支付成功次数为(2+1)/2=1.5 ，前2来自 transaction_id 1 & 2，第二个1来自 transaction_id 4。而连续支付失败的平均次数为在本例中为 1。 最终表格将如下所示：

cus_id |tran_id.|f_or_s |tran_date  |avg_consec_fail|avg_consec_success
  1    |1       |success|2021-01-01 |1              |1.5
  1    |2       |success|2021-01-15 |1              |1.5
  1    |3       |failed |2021-01-30 |1              |1.5
  1    |4       |success|2021-02-15 |1              |1.5

我如何使用R/dplyr实现这一点？

Answer 1

您可以尝试使用rle

数据

df <- read.table(text = "customer_id transaction_id. failed_or_success   transaction_date 
  1         1              success            2021-01-01
  1         2              success            2021-01-15
  1         3              failed             2021-01-30
  1         4              success            2021-02-15", header = TRUE)

代码

df %>%
  mutate(avg_consec_success = mean(rle(failed_or_success)$length[rle(failed_or_success)$values != "failed"]))

  customer_id transaction_id. failed_or_success transaction_date avg_consec_success
1           1               1           success       2021-01-01                1.5
2           1               2           success       2021-01-15                1.5
3           1               3            failed       2021-01-30                1.5
4           1               4           success       2021-02-15                1.5