是否有办法将日期更改为R中协议的一天?

Question

我有一个非常大的受试者数据集，他们接受了两个方案。下面是一个例子df:

df <- data.frame (Subject = c('875','875','875','875','875','875','875','875',
'875','1392','1392','1392','1392','1392','1392',
'1392','1392','1392','1392','1392'), 
StartDate = c('20160915','20160916','20160917','20160918',
'20160926','20160927','20160928','20160929',
'20160930','20160917','20160918','20160919',
'20160920','20160921','20160922','20161005',
'20161006','20161007','20161008','20161009'),
Protocol = c('Training','Training','Training','Training',
'Test','Test','Test','Test','Test','Training',
'Training','Training','Training','Training',
'Training','Test','Test','Test','Test','Test'),
Points = c('14','32','33','50','41','56','41','48','49','34',
'16','25','21','34','28','61','24','45','45','47'),
Attempts = c('24','45','37','69','56','59','53','67','53','55',
'34','36','26','48','44','64','36','62','53','55'))

我想把每个协议的开始日期改为测试日，或者甚至只是在测试日一列中添加。

每个协议都有不规则的日期，所以简单的编号是行不通的。而实际的df有几千行，所以手工计算是一场噩梦。我希望能够比较受试者之间协议的具体日期(即受试者之间培训的第3天等)

我想改变数据帧看起来像:

df2 <- data.frame (Subject = c('875','875','875','875','875','875','875','875',
'875','1392','1392','1392','1392','1392','1392',
'1392','1392','1392','1392','1392'), 
StartDate = c('20160915','20160916','20160917','20160918',
'20160926','20160927','20160928','20160929',
'20160930','20160917','20160918','20160919',
'20160920','20160921','20160922','20161005',
'20161006','20161007','20161008','20161009'),
Day = c('1','2','3','4','1','2','3','4','5','1','2','3',
'4','5','6','1','2','3','4','5'),
Protocol = c('Training','Training','Training','Training',
'Test','Test','Test','Test','Test','Training',
'Training','Training','Training','Training',
'Training','Test','Test','Test','Test','Test'),
Points = c('14','32','33','50','41','56','41','48','49','34',
'16','25','21','34','28','61','24','45','45','47'),
Attempts = c('24','45','37','69','56','59','53','67','53','55',
'34','36','26','48','44','64','36','62','53','55'))

有办法做到这一点吗?或者有人有其他想法吗?

Answer 1

每个协议都有不固定的日期是什么意思?

假设你的数据按StartDate排序。你可以这样做

df %>%
group_by(Subject, Protocol) %>%
mutate(Day = row_number()) %>%
ungroup

# A tibble: 20 × 6
Subject StartDate Protocol Points Attempts   Day
<fct>   <fct>     <fct>    <fct>  <fct>    <int>
1 875     20160915  Training 14     24           1
2 875     20160916  Training 32     45           2
3 875     20160917  Training 33     37           3
4 875     20160918  Training 50     69           4
5 875     20160926  Test     41     56           1
6 875     20160927  Test     56     59           2
7 875     20160928  Test     41     53           3
8 875     20160929  Test     48     67           4
9 875     20160930  Test     49     53           5
10 1392    20160917  Training 34     55           1
11 1392    20160918  Training 16     34           2
12 1392    20160919  Training 25     36           3
13 1392    20160920  Training 21     26           4
14 1392    20160921  Training 34     48           5
15 1392    20160922  Training 28     44           6
16 1392    20161005  Test     61     64           1
17 1392    20161006  Test     24     36           2
18 1392    20161007  Test     45     62           3
19 1392    20161008  Test     45     53           4
20 1392    20161009  Test     47     55           5