Create a vector based on another vector in R?-CodePudding

In my data I have a vector of characters, where one of the characters (in this case, the letter P) is repeated a certain number of times. For example, if I have the following information:

number <- 2
iterations <- 2

and a vector of characters:

df <- data.frame(var = c("P", "a", "b", "P", "d", "a", "k",
                         "P", "e", "q", "s", "P", "d", "v", "i", "j"))

We can see that P is repeated 4 times. Or, to put it another way, P is repeated number * iterations times.

So in my code, I have some function that creates the data frame df, it will create a sequence of strings that always start with P. This happens number of times (in this example 2)... and gets iterated by the iterations number (in this example, also, 2). So that means P appears number * iterations times in my data frame.

Im trying to create a new column that will contain the iteration number. So, in my example, the first two P's would be in iteration 1 and the second two would be in iteration 2. For example, my desired output would look something like:

> df
   var iter
1    P    1
2    a    1
3    b    1
4    P    1
5    d    1
6    a    1
7    k    1
8    P    2
9    e    2
10   q    2
11   s    2
12   P    2
13   d    2
14   v    2
15   i    2
16   j    2

I hope what I said makes sense... I was finding it difficult to accurately word this problem

CodePudding user response：

We may use gl to do this - replace a vector of NAs where the 'var' is 'P' with the gl indexing and then use fill to replace the NA with the non-NA previous value

library(dplyr)
library(tidyr)
df %>% 
   mutate(iter = replace(rep(NA_integer_, n()), var == 'P', 
     as.integer(gl(sum(var == 'P'), 2, sum(var == 'P'))))) %>% 
     fill(iter)

-output

   var iter
1    P    1
2    a    1
3    b    1
4    P    1
5    d    1
6    a    1
7    k    1
8    P    2
9    e    2
10   q    2
11   s    2
12   P    2
13   d    2
14   v    2
15   i    2
16   j    2