查找每个主题的值并将其保存为新表

Question

我有一个具有长格式数据的数据框架，如下所示

ID Frame.No ROI.No Flux.med
01 1        1      78
01 1        2      76
01 2        1      80
01 2        2      80
01 3        1      89
01 3        2      80
27 1        1      60
27 1        2      68
27 4        1      80
27 4        2      89

对于每个"ID“，我希望获得ROI 1和2的第一个和最大的Flux.med，并将所有这些放在一个新的数据want中。如果我有一个只有一个主题(例如ID 01)的数据，我可以使用以下代码识别我需要的Flux.med值：

ROI1.baseline <-  mydata %>%
    filter(ROI.No == "ROI 1" & Frame.No == min(Frame.No))%>%   
  select(Flux.Med)

ROI1.max <-  mydata%>%
     filter(ROI.No == "ROI 1")%>%
  filter (Flux.Med == max(Flux.Med))%>%
  select(Flux.Med)

ROI2.baseline <-  mydata%>%
  filter(ROI.No == "ROI 2" & Frame.No == min(Frame.No))%>%
  select(Flux.Med)

ROI.max <-  mydata%>%
   filter(ROI.No == "ROI 2")%>%
  filter (Flux.Med == max(Flux.Med))%>%
  select(Flux.Med)

但是我需要对每个ID这样做，并将结果保存在一个dataframe中。

我可以用一个for循环来完成这个任务吗？

Answer 1

谢谢你的建议。最后我就是这样做的：

ROI1.baseline <-  ldi_data %>%
  group_by(ID)%>%
  filter(ROI.No == "ROI 1" & Frame.No == min(Frame.No))%>%   ###uses lowest number frame as baseline (not necessarily frame 1 if it was excluded)
  select(Flux.Med)%>%
  dplyr::rename(ROI1_baseline = Flux.Med)%>%
  as.data.frame(ROI1.baseline)

ROI1.max <-  ldi_data%>%
  group_by(ID)%>%
  filter(ROI.No == "ROI 1")%>%
  filter (Flux.Med == max(Flux.Med))%>%
  select(Flux.Med)%>%
  dplyr::rename(ROI1_max = Flux.Med)%>%
  as.data.frame(ROI1.max)

ROI2.baseline <-  ldi_data%>%
  group_by(ID)%>%
  filter(ROI.No == "ROI 2" & Frame.No == min(Frame.No))%>%
  select(Flux.Med)%>%
  dplyr::rename(ROI2_baseline = Flux.Med)%>%
  as.data.frame(ROI2.baseline)

ROI2.max <-  ldi_data%>%
  group_by(ID)%>%
  filter(ROI.No == "ROI 2")%>%
  filter (Flux.Med == max(Flux.Med))%>%
  select(Flux.Med)%>%
  dplyr::rename(ROI2_max = Flux.Med)%>%
  as.data.frame(ROI2.max)

summary <-   Reduce(merge, list(ROI1.baseline, ROI1.max, ROI2.baseline, ROI2.max))

Answer 2

我们可以在每个max和ID中获得第一个和ROI.No值。

library(dplyr)

mydata %>%
  group_by(ID, ROI.No) %>%
  summarise(first_flux = first(Flux.med), 
             max_flux = max(Flux.med))

#    ID ROI.No first_flux max_flux
#  <int>  <int>      <int>    <int>
#1     1      1         78       89
#2     1      2         76       80
#3    27      1         60       80
#4    27      2         68       89

或者使用aggregate：

aggregate(Flux.med~ID + ROI.No, mydata, function(x) c(first = x[1], max = max(x)))

数据

mydata <- structure(list(ID = c(1L, 1L, 1L, 1L, 1L, 1L, 27L, 27L, 27L, 
27L), Frame.No = c(1L, 1L, 2L, 2L, 3L, 3L, 1L, 1L, 4L, 4L), ROI.No = c(1L, 
2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L), Flux.med = c(78L, 76L, 80L, 
80L, 89L, 80L, 60L, 68L, 80L, 89L)), class = "data.frame", row.names = c(NA,-10L))

Answer 3

我们可以使用data.table

library(data.table)
setDT(df1)[, .(first_flux = first(Flux.med), 
         max_flux = max(Flux.med)), .(ID, ROI.No)]

数据

df1 <- structure(list(ID = c(1L, 1L, 1L, 1L, 1L, 1L, 27L, 27L, 27L, 
27L), Frame.No = c(1L, 1L, 2L, 2L, 3L, 3L, 1L, 1L, 4L, 4L), ROI.No = c(1L, 
2L, 1L, 2L, 1L, 2L, 1L, 2L, 1L, 2L), Flux.med = c(78L, 76L, 80L, 
80L, 89L, 80L, 60L, 68L, 80L, 89L)), class = "data.frame", 
row.names = c(NA,-10L))