dplyr group by, carry forward value from previous group to next

Question

Ok this is the over all view of what i\'m trying to achieve with dplyr:

Using dplyr I am making calculations to form new columns.

initial.         


        

Answer 1

It took me a very long time to understand what you are going for: for a single "update", does this work?

library(tidyverse)
library(magrittr)
temp <- df %>% 
  dplyr::mutate(RunID = data.table::rleid(x.long)) %>%
  group_by(RunID) %>% # Don't delete the RunID 
  dplyr::mutate(max.new = max(new.initial.capital)) %>% 
  slice(1) %>%
  arrange(x.long) %>% 
  dplyr::mutate(pass.value = lag(max.new))

df <- left_join(df, temp %>% dplyr::select(x.long, RunID, pass.value)

After this, replace values of initial.capital using pass.value column, according to grouped row_number as you have done above.

I'm not quite sure how to go about it without looping this updating procedure, and I guess if you want to do 10,000 updates like this it will certainly be a bummer. But it will be enable you to "pass" the value to the second red cell as in your picture.