从R中按组求和减去列

Question

Here part of mydataset

df=structure(list(CustomerName = structure(c(1L, 1L, 1L, 2L, 2L, 
2L), .Label = c("x", "y"), class = "factor"), ItemRelation = c(11202L, 
11202L, 11202L, 1L, 1L, 1L), SaleCount = c(214L, 88L, 42L, 214L, 
88L, 42L), DocumentNum = c(137L, 137L, 137L, 3L, 3L, 3L), DocumentYear = c(2018L, 
2018L, 2018L, 2018L, 2018L, 2018L), k = c(114.66667, 114.66667, 
114.66667, 114.66667, 114.66667, 114.66667), m0 = c(31.92, 31.92, 
31.92, 31.92, 31.92, 31.92), Action_Effect = c(82.74667, 82.74667, 
82.74667, 82.74667, 82.74667, 82.74667)), .Names = c("CustomerName", 
"ItemRelation", "SaleCount", "DocumentNum", "DocumentYear", "k", 
"m0", "Action_Effect"), class = "data.frame", row.names = c(NA, 
-6L))

i need for each group CustomerName+ItemRelation+DocumentNum+DocumentYear calculate the sum for salecount and then from this sum substract Action_Effect column.

IE output must be

    df2=structure(list(CustomerName = structure(c(1L, 1L, 1L, 2L, 2L, 
2L), .Label = c("x", "y"), class = "factor"), ItemRelation = c(11202L, 
11202L, 11202L, 1L, 1L, 1L), SaleCount = c(214L, 88L, 42L, 214L, 
88L, 42L), DocumentNum = c(137L, 137L, 137L, 3L, 3L, 3L), DocumentYear = c(2018L, 
2018L, 2018L, 2018L, 2018L, 2018L), X. = c(114.66667, 114.66667, 
114.66667, 114.66667, 114.66667, 114.66667), m0 = c(31.92, 31.92, 
31.92, 31.92, 31.92, 31.92), Action_Effect = c(82.74667, 82.74667, 
82.74667, 82.74667, 82.74667, 82.74667), sum = c(344L, 344L, 
344L, 344L, 344L, 344L), output = c(261.25333, 261.25333, 261.25333, 
261.25333, 261.25333, 261.25333)), .Names = c("CustomerName", 
"ItemRelation", "SaleCount", "DocumentNum", "DocumentYear", "X.", 
"m0", "Action_Effect", "sum", "output"), class = "data.frame", row.names = c(NA, 
-6L))

the long table, so i decided show desired output via dput()

How can i do it?

Answer 1

Your data is a bit weird, as the values are the same for both groups, but this should work:

libary(dplyr)

df %>%
  group_by(CustomerName, ItemRelation, DocumentNum, DocumentYear) %>%
  mutate(test = sum(SaleCount) - Action_Effect)

# A tibble: 6 x 9
# Groups:   CustomerName, ItemRelation, DocumentNum, DocumentYear [2]
  CustomerName ItemRelation SaleCount DocumentNum DocumentYear        k    m0 Action_Effect     test
        <fctr>        <int>     <int>       <int>        <int>    <dbl> <dbl>         <dbl>    <dbl>
1            x        11202       214         137         2018 114.6667 31.92      82.74667 261.2533
2            x        11202        88         137         2018 114.6667 31.92      82.74667 261.2533
3            x        11202        42         137         2018 114.6667 31.92      82.74667 261.2533
4            y            1       214           3         2018 114.6667 31.92      82.74667 261.2533
5            y            1        88           3         2018 114.6667 31.92      82.74667 261.2533
6            y            1        42           3         2018 114.6667 31.92      82.74667 261.2533

---

To add the sum, use

df %>%
  group_by(CustomerName, ItemRelation, DocumentNum, DocumentYear) %>%
  mutate(sum = sum(SaleCount), output = sum(SaleCount) - Action_Effect)

Answer 2

For completeness, adding base and data.table syntax:

base :

df$test <- unlist(by(df, 
    paste(df$CustomerName, df$ItemRelation, df$DocumentNum, df$DocumentYear), 
    function(x) sum(x$SaleCount) - x$Action_Effect))
df

data.table :

library(data.table)
setDT(df)
df[, test2:=sum(SaleCount) - Action_Effect, 
    by=.(CustomerName, ItemRelation, DocumentNum, DocumentYear)][]