Populating one column based on two columns in R-CodePudding

I have the below dataset, and I am trying to create a more meaningful path.

Row#	Session	Click	Page
1	123	Enter	Pg1
2	123	phpbutton	Pg1
3	123	Enter	Pg2
4	123	Enter	Pg3
5	123	Form1	Pg3
6	123	Form2	Pg3
7	123	Form1	Pg3
8	123	Form1	Pg3
9	123	abcbutton	Pg3
10	123	Enter	Pg1
11	123	xyzselect	Pg1
12	123	Enter	Pg4
13	123	Enter	Pg3
14	123	Back	Pg3
15	123	Enter	Pg1

I would like the outcome to look this:

Session	Activity
123	Pg1
123	phpbutton
123	Pg2
123	Pg3
123	Form1
123	Form2
123	Form1
123	abcbutton
123	Pg1
123	xyzselect
123	Pg4
123	Pg3
123	Back
123	Pg1

If the Click column has Enter, then the Activity column should show the Page. But, if the subsequent page is equal to the previous page, then the Activity column should show the value from the Click column. For instance, row# 1 and 2 have the same Page numbers, so I would like the Activity column to show, Pg1, then, phpbutton. But, if the Click column has two or more subsequent same values, as seen in Row# 7 and 8, I would like the Activity column to show just one entry of Form 1.

Thanks a lot.

CodePudding user response：

Try this

df |> group_by(Session) |> 
mutate(Activity = case_when(Click == "Enter" ~ Page , 
lag(Page) == Page ~ Click)) |> select(Session , Activity)

CodePudding user response：

library(dplyr)

dat$activity <- ifelse(dat$click == "Enter" & (lag(dat$page) != dat$page | is.na(lag(dat$page))), dat$page,
       ifelse(lag(dat$page) == dat$page, dat$click, NA))

   row session     click page  activity
1    1     123     Enter  Pg1       Pg1
2    2     123 phpbutton  Pg1 phpbutton
3    3     123     Enter  Pg2       Pg2
4    4     123     Enter  Pg3       Pg3
5    5     123     Form1  Pg3     Form1
6    6     123     Form2  Pg3     Form2
7    7     123     Form1  Pg3     Form1
8    8     123     Form1  Pg3     Form1
9    9     123 abcbutton  Pg3 abcbutton
10  10     123     Enter  Pg1       Pg1
11  11     123 xyzselect  Pg1 xyzselect
12  12     123     Enter  Pg4       Pg4
13  13     123     Enter  Pg3       Pg3
14  14     123      Back  Pg3      Back
15  15     123     Enter  Pg1       Pg1

You could remove duplicate, consecutive rows like this:

dat[cumsum(rle(paste0(dat$session, dat$click, dat$page, dat$activity))$length),]

   row session     click page  activity
1    1     123     Enter  Pg1       Pg1
2    2     123 phpbutton  Pg1 phpbutton
3    3     123     Enter  Pg2       Pg2
4    4     123     Enter  Pg3       Pg3
5    5     123     Form1  Pg3     Form1
6    6     123     Form2  Pg3     Form2
8    8     123     Form1  Pg3     Form1
9    9     123 abcbutton  Pg3 abcbutton
10  10     123     Enter  Pg1       Pg1
11  11     123 xyzselect  Pg1 xyzselect
12  12     123     Enter  Pg4       Pg4
13  13     123     Enter  Pg3       Pg3
14  14     123      Back  Pg3      Back
15  15     123     Enter  Pg1       Pg1

data:

structure(list(row = 1:15, session = c(123L, 123L, 123L, 123L, 
123L, 123L, 123L, 123L, 123L, 123L, 123L, 123L, 123L, 123L, 123L
), click = c("Enter", "phpbutton", "Enter", "Enter", "Form1", 
"Form2", "Form1", "Form1", "abcbutton", "Enter", "xyzselect", 
"Enter", "Enter", "Back", "Enter"), page = c("Pg1", "Pg1", "Pg2", 
"Pg3", "Pg3", "Pg3", "Pg3", "Pg3", "Pg3", "Pg1", "Pg1", "Pg4", 
"Pg3", "Pg3", "Pg1")), row.names = c(NA, -15L), class = "data.frame")