library(data.table)
dt <- data.table(structure(list(helpfulDescriptor = structure(c(1L, 1L, 1L, 1L,1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L,2L, 2L, 2L, 2L, 3L, 3L, 3L, 3L, 3L, 3L, 3L, 3L), .Label = c("ugly_carpet","zestyTomato", "brexit-Vote"), class = "factor"), eventDate = structure(c(15162,15162, 15249, 15249, 15249, 15249, 15250, 15250, 15250, 15250,16868, 16883, 16883, 16883, 16883, 16883, 15414, 15414, 15414,15418, 15418, 16588, 16591, 16591, 15372, 15601, 15601, 16230,16423, 16577, 16577, 16827), class = "Date"), indicator = c(0L,0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L,0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L, 0L)), .Names = c("helpfulDescriptor","eventDate", "indicator"), class = c("data.table", "data.frame")))
dt
helpfulDescriptor eventDate indicator
1: ugly_carpet 2011-07-07 0
2: ugly_carpet 2011-07-07 0
3: ugly_carpet 2011-10-02 0
4: ugly_carpet 2011-10-02 0
5: ugly_carpet 2011-10-02 0
6: ugly_carpet 2011-10-02 0
7: ugly_carpet 2011-10-03 0
8: ugly_carpet 2011-10-03 0
9: ugly_carpet 2011-10-03 0
10: ugly_carpet 2011-10-03 0
11: ugly_carpet 2016-03-08 0
12: ugly_carpet 2016-03-23 0
13: ugly_carpet 2016-03-23 0
14: ugly_carpet 2016-03-23 0
15: ugly_carpet 2016-03-23 0
16: ugly_carpet 2016-03-23 0
17: zestyTomato 2012-03-15 0
18: zestyTomato 2012-03-15 0
19: zestyTomato 2012-03-15 0
20: zestyTomato 2012-03-19 0
21: zestyTomato 2012-03-19 0
22: zestyTomato 2015-06-02 0
23: zestyTomato 2015-06-05 0
24: zestyTomato 2015-06-05 0
25: brexit-Vote 2012-02-02 0
26: brexit-Vote 2012-09-18 0
27: brexit-Vote 2012-09-18 0
28: brexit-Vote 2014-06-09 0
29: brexit-Vote 2014-12-19 0
30: brexit-Vote 2015-05-22 0
31: brexit-Vote 2015-05-22 0
32: brexit-Vote 2016-01-27 0
我正在努力使用data.table来识别共享由helpfulDescriptor分组的相同，最近，（最大）日期的所有行。
dt[eventDate == max(eventDate), indicator := 1L, by = c('helpfulDescriptor')]
这仅将指标设置为1表示整个data.table的最大日期，而不是分组...
我的其他尝试使用tail，setkey（）,. SD，ect ......失败。
正确的解决方案是将指标设置为1，最后（在此排列列表中）最近（日期）5行ugly_carpet，2 2015-06-05 zestyTomato，以及1个最终或最近的brexitvotes。 / p>
谢谢。