Create column based on multiple column and row criteria in R-CodePudding

I am after some help with a bit of complex dataframe management/column creation.

I have a data frame that looks something like this:

dat <- data.frame(Target = sort(rep(c("1", "2", "3", "4"), times = 5)),
                  targ1frames_x = c(0.42, 0.46, 0.50, 0.60, 0.86, 0.32, 0.56, 
                                    0.89, 0.98, 0.86, 0.57, 0.79, 0.52, 0.62, 
                                    0.55, 0.12, 0.35, 0.50, 0.48, 0.45),
                  targ1frames_y = c(0.69, 0.63, 0.74, 0.81, 0.12, 0.67, 0.54, 
                                    0.30, 0.25, 0.18, 0.63, 0.63, 0.79, 0.81, 
                                    0.96, 0.90, 0.75, 0.74, 0.81, 0.76),
                  targ1frames_time = rep(c(0.00006, 0.18, 0.27, 0.35, 0.43), times = 4))

     Target targ1frames_x targ1frames_y targ1frames_time
1       1          0.42          0.69          0.00006
2       1          0.46          0.63          0.18000
3       1          0.50          0.74          0.27000
4       1          0.60          0.81          0.35000
5       1          0.86          0.12          0.43000
6       2          0.32          0.67          0.00006
7       2          0.56          0.54          0.18000
8       2          0.89          0.30          0.27000
9       2          0.98          0.25          0.35000
10      2          0.86          0.18          0.43000
11      3          0.57          0.63          0.00006
12      3          0.79          0.63          0.18000
13      3          0.52          0.79          0.27000
14      3          0.62          0.81          0.35000
15      3          0.55          0.96          0.43000
16      4          0.12          0.90          0.00006
17      4          0.35          0.75          0.18000
18      4          0.50          0.74          0.27000
19      4          0.48          0.81          0.35000
20      4          0.45          0.76          0.43000

This is from an experiment in which participants had to click on-screen targets with a computer mouse. The Target column indicates which target was clicked, targ1frames_x shows the x coordinates of the cursor on each frame; targ1frames_y shows the y coordinates on each frame; and targ1frames_time indicates the time of each frame change.

What I want to work out is the time of the frame change when the mouse coordinates from either x or y columns changed between rows by >= 0.2, and get the single time for each target when this occurs into a new column/data frame. I.e., I want to end up with:

    Target targ1_initTime
1      1           0.43000
2      2           0.18000
3      3           0.18000
4      4           0.18000

Note that for the simplicity of providing a reproducible example, each target here has the same time values, which is not the case in my actual dataset. I am trying to write code that can be flexible to produce the time at which the x or y coordinate changes across rows (within the x or y coordinate columns) by -0.2 for each target.

I hope this makes sense, I would be very grateful for any advice - if anything needs further clarification, please let me know in the comments.

CodePudding user response：

We could do it this way. Here demonstrated for targ1frames_x (for targ1frames_y the code needs to be adapted):

library(dplyr)

dat %>% 
  group_by(Target) %>% 
  mutate(diff_x = targ1frames_x -  lag(targ1frames_x)) %>% 
  filter(diff_x >= 0.2) %>% 
  slice(1) %>% 
  select(Target, targ1frames_time)

  Target targ1frames_time
  <chr>             <dbl>
1 1                  0.43
2 2                  0.18
3 3                  0.18
4 4                  0.18

CodePudding user response：

dat %>%  
  mutate(diff_row_x = targ1frames_x  - lag(targ1frames_x , default = first(targ1frames_x )))%>%
  mutate(diff_row_y = targ1frames_y  - lag(targ1frames_y , default = first(targ1frames_y )))%>%
  filter(!(diff_row_x & diff_row_y <=0.2))

Now I understand your question, this code is more appropriate than my previous solution