About, creating a new column in data frame as a function of other columns

[prateek26394]

I want to create a new column in a data frame as a average of values from an already existing column. The average should be in iterative order like

[ablack3]

How about this?

df <- data.frame(col_a = c(69, 77, 55, 58, 15, 42, 94, 54, 87, 96, 31, 75, 4, 31, 23, 37, 41, 6, 28))

library(dplyr)

# This is almost what you want
mutate(df, col_b = (cumsum(col_a) - lag(cumsum(col_a), n = 10))/10)
#>    col_a col_b
#> 1     69    NA
#> 2     77    NA
#> 3     55    NA
#> 4     58    NA
#> 5     15    NA
#> 6     42    NA
#> 7     94    NA
#> 8     54    NA
#> 9     87    NA
#> 10    96    NA
#> 11    31  60.9
#> 12    75  60.7
#> 13     4  55.6
#> 14    31  52.9
#> 15    23  53.7
#> 16    37  53.2
#> 17    41  47.9
#> 18     6  43.1
#> 19    28  37.2

# This should give exactly what you want
mutate(df, col_b = ifelse(is.na(lag(col_a, 10)) & !is.na(lag(col_a, 9)), cummean(col_a), (cumsum(col_a) - lag(cumsum(col_a), n = 10))/10))
#>    col_a col_b
#> 1     69    NA
#> 2     77    NA
#> 3     55    NA
#> 4     58    NA
#> 5     15    NA
#> 6     42    NA
#> 7     94    NA
#> 8     54    NA
#> 9     87    NA
#> 10    96  64.7
#> 11    31  60.9
#> 12    75  60.7
#> 13     4  55.6
#> 14    31  52.9
#> 15    23  53.7
#> 16    37  53.2
#> 17    41  47.9
#> 18     6  43.1
#> 19    28  37.2

^{Created on 2018-11-22 by the reprex package (v0.2.1)}

[prateek26394]

@ablack3

mutate(df, col_b = ifelse(is.na(lag(col_a, 10)) & !is.na(lag(col_a, 9)), cummean(col_a), (cumsum(col_a) - lag(cumsum(col_a), n = 10))/10))

Thank You for this solution:hugs: