我的输入是前三列：ID COLLECTED_DATE Hyperkalemia
我的输出是最后三列：met_criteria my_row next_1_time_interval
Met_criteria =高钾血症== 3 | 4
My_row =在3或4值（在同一ID内）后，高钾血症= = 1的行
next_1_time_interval == 3或4值与时间1值之间的时间（在同一ID内）
> ROW ID COLLECTED_DATE Hyperkalemia met_criteria my_row next_1_time_interval
1 123 1 5/25/08 3:30 0 NA NA NA
2 124 1 5/26/08 3:30 0 NA NA NA
3 125 1 5/27/08 4:00 2 NA NA NA
4 126 1 5/28/08 3:45 2 NA NA NA
5 127 1 5/29/08 3:40 3 TRUE 248 5.590500e+04
6 128 1 5/29/08 6:40 2 NA NA NA
7 129 1 5/29/08 20:15 2 NA NA NA
8 144 1 6/13/08 4:00 0 NA NA NA
9 145 1 6/14/08 4:10 0 NA NA NA
10 146 1 6/15/08 4:45 4 TRUE 248 5.549592e+04
11 147 1 6/15/08 6:30 0 NA NA NA
12 246 1 12/18/12 10:46 0 NA NA NA
13 247 1 1/14/14 11:30 0 NA NA NA
14 248 1 10/14/14 12:40 1 NA NA NA
15 249 1 2/4/15 14:27 0 NA NA NA
16 633 2 9/12/14 4:30 0 NA NA NA
17 634 2 9/29/14 8:57 0 NA NA NA
18 635 3 12/6/11 2:50 0 NA NA NA
19 636 3 12/6/11 10:45 0 NA NA NA
20 637 3 12/6/11 17:50 0 NA NA NA
21 704 3 4/18/12 5:00 0 NA NA NA
22 705 3 4/18/12 5:00 0 NA NA NA
23 706 3 4/18/12 16:30 3 TRUE 708 8.333333e-01
24 707 3 4/18/12 16:30 4 TRUE 708 8.333333e-01
25 708 3 4/18/12 17:20 1 NA NA NA
26 768 3 5/10/12 3:35 0 NA NA NA
27 769 3 5/11/12 3:20 0 NA NA NA
28 770 3 5/12/12 4:00 4 TRUE NA NA
29 771 3 5/12/12 5:45 0 NA NA NA
30 772 3 5/13/12 4:00 0 NA NA NA
31 773 3 5/14/12 5:30 0 NA NA NA
32 774 3 5/15/12 4:00 4 TRUE NA NA
33 775 3 5/15/12 5:15 0 NA NA NA
34 776 3 5/16/12 8:34 0 NA NA NA
35 777 3 5/17/12 7:10 0 NA NA NA
36 778 3 5/18/12 7:10 0 NA NA NA
37 779 3 5/28/12 7:26 0 NA NA NA
38 780 3 6/20/12 9:46 0 NA NA NA
39 781 3 8/1/12 13:10 0 NA NA NA
40 782 3 11/14/12 13:34 0 NA NA NA
41 783 3 11/13/13 9:35 0 NA NA NA
42 784 3 6/5/15 9:31 0 NA NA NA
列出的输出是所需的输出。
要查看的行是：127,146（3或4的高钾血症）和248（3或4值后的1的高钾血症），每行属于一个人的ID。
另外：行706,707 :(高钾血症中的3或4值 对于相同的ID
，在3或4值（高钾血症）之后的708：1值 pots_1 <- arrange(pots_1, MRN, COLLECTED_DATE)
pots_1$Hyperkalemia[is.na(pots_1$Hyperkalemia)] <- 0
pots_1$met_criteria <- NA
pots_1$next_1_time_interval <- NA
pots_1$my_row <- NA
# data.frame = pots
# for every patient (sorted by mrn and collected date)
for (mrn in unique(pots_1$MRN)){
# for each row for each patient (sorted by collected date)
for (i in 1:length(which(pots_1$MRN == mrn))) {
# number of rows for this mrn
mrn_max_row <- length(which(pots_1$MRN == mrn))
# if Hyperkalemia = 3 or 4
if(pots_1$Hyperkalemia[i] == 3 | pots_1$Hyperkalemia[i] == 4){
my_start_row <- i
my_row <- i
pots_1$met_criteria[i] <- TRUE
#for every row after 3 or 4 until I get to Hyperkalemia = 1
while (pots_1$Hyperkalemia[my_row] != 1){
my_row <- my_row+1
if (my_row > mrn_max_row) {
break
}
} # ends while
if (pots_1$Hyperkalemia[my_row]) {
pots_1$my_row[i] <- my_row
pots_1$next_1_time_interval[my_start_row] <- difftime(pots_1$COLLECTED_DATE[my_row], pots_1$COLLECTED_DATE[my_start_row], units = "hours")
}
# break only marks the first one if break is here
} # ends if
} # ends loop through mrn result rows
} # ends loop through mrns
此代码可正常工作，但仅适用于第一个ID。我已经尝试了许多方法来使这个工作用于整个ID集并且已经获得了许多有趣的错误:)比如计算下一个值而不管ID。
有人可以帮我解决整个ID集的问题吗？
这是用R.写的。我意识到循环和while语句在R中不典型，但是我只有7500行，无论循环如何都运行得非常快。
答案 0 :(得分：1)
以下是使用data.table
library(data.table)
setDT(x)[,COLLECTED_DATE:=
as.POSIXct(COLLECTED_DATE,format="%m/%d/%y %H:%M")]
简单的部分：
x[,met_criteria:=Hyperkalemia %in% 3:4]
现在，定义my_row看起来有些复杂。如果您想要分析其来源，请检查以前的编辑。
x[,my_row:=zoo::na.locf(NA^(Hyperkalemia!=1)*.I,
fromLast=T,na.rm=F)*NA^(!met_criteria),by=ID]
现在next_1_time_interval非常容易：
x[,next_1_time_interval:=difftime(COLLECTED_DATE[my_row],
COLLECTED_DATE,
units="hours")]
> x
ROW ID COLLECTED_DATE Hyperkalemia met_criteria my_row next_1_time_interval
1: 123 1 2008-05-25 03:30:00 0 FALSE NA NA hours
2: 124 1 2008-05-26 03:30:00 0 FALSE NA NA hours
3: 125 1 2008-05-27 04:00:00 2 FALSE NA NA hours
4: 126 1 2008-05-28 03:45:00 2 FALSE NA NA hours
5: 127 1 2008-05-29 03:40:00 3 TRUE 14 5.590500e+04 hours
6: 128 1 2008-05-29 06:40:00 2 FALSE NA NA hours
7: 129 1 2008-05-29 20:15:00 2 FALSE NA NA hours
8: 144 1 2008-06-13 04:00:00 0 FALSE NA NA hours
9: 145 1 2008-06-14 04:10:00 0 FALSE NA NA hours
10: 146 1 2008-06-15 04:45:00 4 TRUE 14 5.549592e+04 hours
11: 147 1 2008-06-15 06:30:00 0 FALSE NA NA hours
12: 246 1 2012-12-18 10:46:00 0 FALSE NA NA hours
13: 247 1 2014-01-14 11:30:00 0 FALSE NA NA hours
14: 248 1 2014-10-14 12:40:00 1 FALSE NA NA hours
15: 249 1 2015-02-04 14:27:00 0 FALSE NA NA hours
16: 633 2 2014-09-12 04:30:00 0 FALSE NA NA hours
17: 634 2 2014-09-29 08:57:00 0 FALSE NA NA hours
18: 635 3 2011-12-06 02:50:00 0 FALSE NA NA hours
19: 636 3 2011-12-06 10:45:00 0 FALSE NA NA hours
20: 637 3 2011-12-06 17:50:00 0 FALSE NA NA hours
21: 704 3 2012-04-18 05:00:00 0 FALSE NA NA hours
22: 705 3 2012-04-18 05:00:00 0 FALSE NA NA hours
23: 706 3 2012-04-18 16:30:00 3 TRUE 25 8.333333e-01 hours
24: 707 3 2012-04-18 16:30:00 4 TRUE 25 8.333333e-01 hours
25: 708 3 2012-04-18 17:20:00 1 FALSE NA NA hours
26: 768 3 2012-05-10 03:35:00 0 FALSE NA NA hours
27: 769 3 2012-05-11 03:20:00 0 FALSE NA NA hours
28: 770 3 2012-05-12 04:00:00 4 TRUE NA NA hours
29: 771 3 2012-05-12 05:45:00 0 FALSE NA NA hours
30: 772 3 2012-05-13 04:00:00 0 FALSE NA NA hours
31: 773 3 2012-05-14 05:30:00 0 FALSE NA NA hours
32: 774 3 2012-05-15 04:00:00 4 TRUE NA NA hours
33: 775 3 2012-05-15 05:15:00 0 FALSE NA NA hours
34: 776 3 2012-05-16 08:34:00 0 FALSE NA NA hours
35: 777 3 2012-05-17 07:10:00 0 FALSE NA NA hours
36: 778 3 2012-05-18 07:10:00 0 FALSE NA NA hours
37: 779 3 2012-05-28 07:26:00 0 FALSE NA NA hours
38: 780 3 2012-06-20 09:46:00 0 FALSE NA NA hours
39: 781 3 2012-08-01 13:10:00 0 FALSE NA NA hours
40: 782 3 2012-11-14 13:34:00 0 FALSE NA NA hours
41: 783 3 2013-11-13 09:35:00 0 FALSE NA NA hours
42: 784 3 2015-06-05 09:31:00 0 FALSE NA NA hours
ROW ID COLLECTED_DATE Hyperkalemia met_criteria my_row next_1_time_interval