在R中的dplyr中使用带有mutate和case_的shift:未按预期工作
我的数据显示了瞳孔大小的变化。当值为-1时，表示闪烁。我已经编写了一些代码来检测闪烁启动和偏移，但是我在使用在R中的dplyr中使用带有mutate和case_的shift:未按预期工作,r,dataframe,dplyr,mutate,R,Dataframe,Dplyr,Mutate,我的数据显示了瞳孔大小的变化。当值为-1时，表示闪烁。我已经编写了一些代码来检测闪烁启动和偏移，但是我在使用shift函数时遇到了一些问题 我的数据样本： library(dplyr) DataFrame<-structure(list(Pupil_Avg = c(7.174, 6.6910005, 6.518, 2.461, 2.182, 1.942, 1.942, -1, -1, -1, -1, -1, -1,
shiftlibrary(dplyr)
DataFrame<-structure(list(Pupil_Avg = c(7.174, 6.6910005, 6.518, 2.461,
2.182, 1.942, 1.942, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1,
-1, -1, -1, -1, -1, -1, -1, 1.487, -1, -1, -1, -1, 2.202, 2.202,
2.281, 2.344)), row.names = c(NA, -481L), class = c("tbl_df",
"tbl", "data.frame"))
DataFrame$BLINK_IDENTIFICATION <- ""
# set an arbitrary decrease value in pupil size
Pupil_Constriction = 3
DataFrame<-DataFrame %>%
# removed the columns below as they're not necessary to work on the problem.
#group_by(StimulusName, Name, StimuliBlock) %>%
# if there is a reduction in pupil size of the value in Pupil_Constriction in the current row add "Blink Onset"
mutate(BLINK_IDENTIFICATION = case_when((DataFrame$Pupil_Avg <= (shift(DataFrame$Pupil_Avg, 1L, type="lag")-Pupil_Constriction)) ~ "Blink Onset",
# The next line of code is supposed to check the last row and if "Blink Onset" is contained in the previous row in BLINK_IDENTIFICATION AND the current Pupil_Avg value is NOT equal to -1, then the write "Blink Onset" to the current BLINK_IDENTIFICATION row
( (shift(DataFrame$BLINK_IDENTIFICATION, 1L, type="lag")=="Blink Onset") & (DataFrame$Pupil_Avg != -1) ) ~ "Blink Onset",
# the next line of code write "Blink Offset" if previous row was -1, current is greater than -1, and the next row is NOT -1
( (shift(DataFrame$Pupil_Avg, 1L, type="lag")==-1) & (DataFrame$Pupil_Avg >-1) & (shift(DataFrame$Pupil_Avg, 1L, type="lead")!=-1)) ~ "Blink Offset",
# the next line write "Eye Closed" if current row equals -1
(DataFrame$Pupil_Avg==-1) ~ "Eye Closed"))
库（dplyr）
DataFrame3您的案例陈述的选项可以在dplyr中完成：
DataFrame <- DataFrame %>%
mutate(BLINK_IDENTIFICATION = case_when(Pupil_Avg == -1 ~ "Eye Closed",
Pupil_Avg <= lag(Pupil_Avg) - Pupil_Constriction ~ "Blink Onset",
lag(Pupil_Avg) == -1 & Pupil_Avg > -1 & lead(Pupil_Avg) != -1 ~ "Blink Offset",
TRUE ~ ""))
# A tibble: 481 x 2
Pupil_Avg BLINK_IDENTIFICATION
<dbl> <chr>
1 7.17 ""
2 6.69 ""
3 6.52 ""
4 2.46 Blink Onset
5 2.18 ""
6 1.94 ""
7 1.94 ""
8 -1 Eye Closed
9 -1 Eye Closed
10 -1 Eye Closed
# ... with 471 more rows
但是你也可以在for循环中做任何事情。正如您在数据末尾看到的，有些空白处没有填写。在这里，你需要定义你想用它们做什么。保持原样，或填写。3使用dplyr可以完成案例陈述的选项：
DataFrame <- DataFrame %>%
mutate(BLINK_IDENTIFICATION = case_when(Pupil_Avg == -1 ~ "Eye Closed",
Pupil_Avg <= lag(Pupil_Avg) - Pupil_Constriction ~ "Blink Onset",
lag(Pupil_Avg) == -1 & Pupil_Avg > -1 & lead(Pupil_Avg) != -1 ~ "Blink Offset",
TRUE ~ ""))
# A tibble: 481 x 2
Pupil_Avg BLINK_IDENTIFICATION
<dbl> <chr>
1 7.17 ""
2 6.69 ""
3 6.52 ""
4 2.46 Blink Onset
5 2.18 ""
6 1.94 ""
7 1.94 ""
8 -1 Eye Closed
9 -1 Eye Closed
10 -1 Eye Closed
# ... with 471 more rows
但是你也可以在for循环中做任何事情。正如您在数据末尾看到的，有些空白处没有填写。在这里，你需要定义你想用它们做什么。保持原样，或填写。shift
来自data.table。dplyr还具有lag
和lead
功能。你为什么把这两个包裹混在一起？@phiver我不知道。你能推荐一个可以使用的代码版本吗？@mt1022我在输入管道代码之前创建了BLINK_标识。我运行了这行DataFrame$BLINK\u IDENTIFICATION@Docconcoct，对不起。我没听到那句话。将很快删除此消息。shift
来自data.table。dplyr还具有lag
和lead
功能。你为什么把这两个包裹混在一起？@phiver我不知道。你能推荐一个可以使用的代码版本吗？@mt1022我在输入管道代码之前创建了BLINK_标识。我运行了这行DataFrame$BLINK\u IDENTIFICATION@Docconcoct，对不起。我没听到那句话。将很快删除此邮件。谢谢。我会在早上检查你的解决方案。我一直在避免使用for循环，因为我处理的是数百万行的海量数据集。我觉得%>%更快。谢谢你的帮助。我会在早上检查你的解决方案。我一直在避免使用for循环，因为我处理的是数百万行的海量数据集。我觉得%>%更快。谢谢你的帮助。