我有一个由研究设备创建的非常混乱的数据集。此数据显示每几毫秒(“时间”)的生理测量(“生理”)。输出列出了一些用户消息,例如试验何时开始(“ START_TRIAL n”),试验何时结束(“ STOP_TRIAL”)以及研究人员可能感兴趣的其他随机事件。有时,“ START_TRIAL n”消息会连续重复,有时没有消息时,在原本为空白单元格的地方会留下一个简单的“ 0”。
我希望创建一个新列,以表示当前案件属于哪个审判。 (请参见下面的示例数据)。
是否可以使用dplyr和mutate做到这一点?我想知道是否可能需要执行if-then语句来更改每种情况下新列的值,但是肯定有更优雅的解决方案吗? (预先感谢您帮助这个新手!)
Time Physio Cond
1 34 START_TRIAL 1
2 33 0
3 25 RANDOM_MSG
4 43 STOP_TRIAL
5 27 START_TRIAL 2
6 54 START_TRIAL 2
7 32 0
8 54 RANDOM_MSG
9 23 STOP_TRIAL
structure(list(Time = 1:9, Physio = c(34L, 33L, 25L, 43L, 27L,
54L, 32L, 54L, 23L), Cond = structure(c(4L, 2L, 3L, 6L, 5L, 5L,
2L, 3L, 6L), .Label = c("", "0", "RANDOM_MSG", "START_TRIAL 1",
"START_TRIAL 2", "STOP_TRIAL"), class = "factor")), .Names = c("Time",
"Physio", "Cond"), row.names = c(NA, 9L), class = "data.frame")
进入
Time Physio Trial Cond
1 34 1 START_TRIAL 1
2 33 1 0
3 25 1 RANDOM_MSG
4 43 1 STOP_TRIAL
5 27 2 START_TRIAL 2
6 54 2 START_TRIAL 2
7 32 2 0
8 54 2 RANDOM_MSG
9 23 2 STOP_TRIAL
structure(list(Time = 1:9, Physio = c(34L, 33L, 25L, 43L, 27L,
54L, 32L, 54L, 23L), Trial = c(1L, 1L, 1L, 1L, 2L, 2L, 2L, 2L,
2L), Cond = structure(c(4L, 2L, 3L, 6L, 5L, 5L, 2L, 3L, 6L), .Label = c("",
"0", "RANDOM_MSG", "START_TRIAL 1", "START_TRIAL 2", "STOP_TRIAL"
), class = "factor")), .Names = c("Time", "Physio", "Trial",
"Cond"), row.names = c(NA, 9L), class = "data.frame")
答案 0 :(得分:3)
一种选择是用grep
标识“ START_TRIAL”,执行match
以获取索引,并fill
将NA元素与先前的非NA相邻元素>
library(dplyr)
library(tidyr)
df1 %>%
mutate(Trial = match(PhysioCond, unique(grep("START_TRIAL",
PhysioCond, value = TRUE)))) %>%
fill(Trial)
# Time PhysioCond Trial
#1 34 START_TRIAL 1 1
#2 33 0 1
#3 25 RANDOM_MSG 1
#4 43 STOP_TRIAL 1
#5 27 START_TRIAL 2 2
#6 54 START_TRIAL 2 2
#7 32 0 2
#8 54 RANDOM_MSG 2
#9 23 STOP_TRIAL 2
注意:不清楚列名,但是逻辑应该工作正常
df1 <- structure(list(Time = c(34L, 33L, 25L, 43L, 27L, 54L, 32L, 54L,
23L), PhysioCond = c("START_TRIAL 1", "0", "RANDOM_MSG", "STOP_TRIAL",
"START_TRIAL 2", "START_TRIAL 2", "0", "RANDOM_MSG", "STOP_TRIAL"
)), class = "data.frame", row.names = c("1", "2", "3", "4", "5",
"6", "7", "8", "9"))