如何在 tibble 列中找到最长的重复序列（跟进问题）？

Question

这是我在此处发布的问题的后续问题：

作为输出，我还需要一列（年）：

library(tibble)
library(purrr)

my_tbl <- tribble(
  ~year, ~event_id, ~winner_id,
   2011,      "A",     4322,
   2012,      "A",     4322,
   2013,      "A",     4322,
   2014,      "A",     5478,
   2015,      "A",     4322,
   2011,      "B",     4322,
   2012,      "B",     7893,
   2013,      "B",     7893,
   2014,      "B",     2365,
   2015,      "B",     3407,
   2011,      "C",     5556,
   2012,      "C",     5556,
   2013,      "C",     1238,
   2014,      "C",     2391,
   2015,      "C",     2391,
   2011,      "D",     4219,
   2012,      "D",     7623,
   2013,      "D",     8003,
   2014,      "D",     2851,
   2015,      "D",     0418
)

 results_summary_tbl <- tribble(
    ~event_id, ~most_wins_in_a_row, ~number_of_winners, ~winners,                              ~years,
    "A",       3,                  1,                   "4322",                               "4322 = (2011, 2012, 2013)",
    "C",       2,                  2,                   "5556 , 2391",                        "5556 = (2011, 2012), 2391 = (2014, 2015)",
    "B",       2,                  1,                   "7893",                               "7893 = (2012, 2013)",
    "D",       1,                  5,                   "4219 , 7623 , 8003 , 2851 , 0418",   "4219 = (2011), 7623 = (2012), 8003 = (2013), 2851 = (2014), 0418 = (2015)"
 )

谢谢，

Answer 1

继续@tmfmnk 上次post的方法，我们可以通过

实现预期的输出

library(dplyr)

my_tbl %>%
   add_count(event_id, rleid = cumsum(winner_id != lag(winner_id, default = first(winner_id)))) %>%
   group_by(event_id) %>%
   mutate(most_wins_in_a_row = max(n),
          number_of_winners = n_distinct(winner_id[n == max(n)]),
          winners = paste0(unique(winner_id[n == max(n)]), collapse = ",")) %>%
    group_by(event_id, winner_id) %>%
    mutate(year = paste0(first(winner_id), "= (", toString(year), ")")) %>%
    group_by(event_id) %>%
    mutate(year = toString(unique(year))) %>%
    slice(1L)

如果您将 pull(year) 添加到上述链中，您将看到输出为

#[1] "4322= (2011, 2012, 2013, 2015), 5478= (2014)"                       
#[2] "4322= (2011), 7893= (2012, 2013), 2365= (2014), 3407= (2015)"       
#[3] "5556= (2011, 2012), 1238= (2013), 2391= (2014, 2015)"               
#[4] "4219= (2011), 7623= (2012), 8003= (2013), 2851= (2014), 418= (2015)"

Answer 2

使用 data.table 您可以查看哪些行具有最高 rowid(rleid(winner_id))，即连续重复的最高数量，并使用它来创建中间 table。从那里计算更容易。

library(data.table)
setDT(my_tbl)

my_tbl[, ro := rowid(rleid(winner_id, event_id))]
most_wins <- my_tbl[my_tbl[, .(ro = max(ro)), event_id], on = .(ro, event_id)]
most_wins[, year := Map(function(x, y) toString(x - y:1 + 1), year, ro)]

most_wins[, .(most_wins_in_a_row = ro[1],
              number_of_winners = .N,
              winners = toString(winner_id),
              years = toString(paste0(winner_id, ' = (', year, ')')))
          , by = event_id]

#    event_id most_wins_in_a_row number_of_winners                     winners
# 1:        A                  3                 1                        4322
# 2:        B                  2                 1                        7893
# 3:        C                  2                 2                  5556, 2391
# 4:        D                  1                 5 4219, 7623, 8003, 2851, 418
#                                                                       years
# 1:                                                4322 = (2011, 2012, 2013)
# 2:                                                      7893 = (2012, 2013)
# 3:                                 5556 = (2011, 2012), 2391 = (2014, 2015)
# 4: 4219 = (2011), 7623 = (2012), 8003 = (2013), 2851 = (2014), 418 = (2015)

如何在 tibble 列中找到最长的重复序列（跟进问题）？

How to find the longest duplicate sequence in a tibble column (follow up question)?

r

run-length-encoding

tibble