合并 group_by() 之间的连续间隔

Question

我对某些数据有疑问，正在寻求您的帮助。 这是包含数千艘船只的较大数据集的子集：

subset <- tibble::tribble(
           ~cfr, ~vessel_name, ~reg_port, ~event_start_date, ~event_end_date, ~length, ~tonnage, ~power,                     ~gear,         ~gear_cat,      ~gear_type,
 "FRA000859026",   "PATOLISA",      "ST",      "1995-04-10",    "1996-09-06",    7.25,     3.02,    110, "Set gillnets (anchored)", "Entangling nets", "Passive gears",
 "FRA000859026",   "PATOLISA",      "ST",      "1996-09-07",    "1996-12-31",    7.25,     3.02,     85, "Set gillnets (anchored)", "Entangling nets", "Passive gears",
 "FRA000859026",   "PATOLISA",      "ST",      "1997-01-01",    "1999-12-01",    7.25,     3.02,     85, "Set gillnets (anchored)", "Entangling nets", "Passive gears",
 "FRA000859026",   "PATOLISA",      "ST",      "1999-12-02",    "2000-02-03",    7.25,     3.02,     85, "Set gillnets (anchored)", "Entangling nets", "Passive gears",
 "FRA000859026",   "PATOLISA",      "ST",      "2000-02-04",    "2001-06-10",    7.25,     3.02,    110, "Set gillnets (anchored)", "Entangling nets", "Passive gears",
 "FRA000859026",   "PATOLISA",      "ST",      "2001-06-11",    "2001-07-23",    7.25,     3.02,    110, "Set gillnets (anchored)", "Entangling nets", "Passive gears",
 "FRA000859026",   "PATOLISA",      "ST",      "2001-07-24",    "2002-12-31",    7.25,     3.02,    110, "Set gillnets (anchored)", "Entangling nets", "Passive gears",
 "FRA000859026",   "PATOLISA",      "ST",      "2003-01-01",    "2004-03-10",    7.25,     3.02,    110, "Set gillnets (anchored)", "Entangling nets", "Passive gears"
 )

我想简化这些数据，以便为每组相似的cfr、vessel_name、reg_port、length、tonnage、 power、gear、gear_cat 和 gear_type。我的预期结果如下所示：

~cfr, ~vessel_name, ~reg_port, ~event_start_date, ~event_end_date, ~length, ~tonnage, ~power,                     ~gear,         ~gear_cat,      ~gear_type,
 "FRA000859026",   "PATOLISA",      "ST",      "1995-04-10",    "1996-09-06",    7.25,     3.02,    110, "Set gillnets (anchored)", "Entangling nets", "Passive gears",
 "FRA000859026",   "PATOLISA",      "ST",      "1996-09-07",    "2000-02-03",    7.25,     3.02,     85, "Set gillnets (anchored)", "Entangling nets", "Passive gears",
 "FRA000859026",   "PATOLISA",      "ST",      "2000-02-04",    "2004-03-10",    7.25,     3.02,    110, "Set gillnets (anchored)", "Entangling nets", "Passive gears"
 )

但是，无论我尝试什么，我的结果总是将所有记录与 power = 110 结合起来，即使两者之间有一个间隔，power = 85。

特别是，我尝试了几件事，但没有按预期工作：

1. group_by() 和 mutate()

subset %>%
  group_by(cfr, vessel_name, reg_port, length, tonnage, power, gear, gear_cat, gear_type) %>%
  mutate(event_start_date = min(event_start_date), #Find oldest date for group
           event_end_date = max(event_end_date)) %>% #Find most recent date for group
    ungroup() %>%
    distinct()

==> 所有 power = 110 的记录都被认为是相似的，这会创建两个愚蠢的重叠记录，一个是 power = 110 从 1995-04-10 到 2004-03-10，另一个是 power = 85 从 1996- 09-07 至 2000-02-03

2。 cur_group_id()

所以我认为为每个组创建一个连续的id 是有意义的，这样我就可以执行类似的 mutate() 但这次在我的分组中使用id。我尝试使用 cur_group_id()，但结果是一样的：所有 power = 110 的记录都被认为是相似的，即使我确实想将两个时间组分开。

subset %>%
  group_by(cfr, vessel_name, reg_port, length, tonnage, power, gear, gear_cat, gear_type) %>%
  mutate(id = cur_group_id()) %>%
  ungroup() %>%
  group_by(id) %>%
  mutate(event_start_date = min(event_start_date), #Find oldest date for group
           event_end_date = max(event_end_date)) %>% #Find most recent date for group
    ungroup() %>%
    distinct()

如何确保不会发生这种情况并获得预期的输出，即按组合并的间隔但考虑时间变化？这可能很简单，但我无法理解它......

谢谢！

Answer 1

您可以使用data.table 和rleid：

library(data.table)
setDT(subset)


subset[,.(event_start_date = min(event_start_date),
          event_end_date   = max(event_end_date)),
       by=.(rleid(cfr, vessel_name, reg_port, length, tonnage, power,gear, gear_cat, gear_type ),
            cfr,
            vessel_name,
            reg_port,
            length,
            tonnage,
            power,
            gear,
            gear_cat,
            gear_type)]

            cfr vessel_name reg_port length tonnage power rleid                    gear        gear_cat     gear_type event_start_date event_end_date
1: FRA000859026    PATOLISA       ST   7.25    3.02   110     1 Set gillnets (anchored) Entangling nets Passive gears       1995-04-10     1996-09-06
2: FRA000859026    PATOLISA       ST   7.25    3.02    85     2 Set gillnets (anchored) Entangling nets Passive gears       1996-09-07     2000-02-03
3: FRA000859026    PATOLISA       ST   7.25    3.02   110     3 Set gillnets (anchored) Entangling nets Passive gears       2000-02-04     2004-03-10