Selectively apply custom function based on criteria - r

Question

I am working with this dataframe:
structure(list(year = c("2012", "2016", "2012", "2016"), month = c("12",
"12", "12", "12"), company = c("ALSN", "ALSN", "DAN", "DAN"),
Revenue = c(2141.8, 1840.2, 7224, 5826), `Cost of Goods Sold` = c(1187.5,
976, 6250, 4982), `Gross Profit` = c(954.3, 864.2, 974, 844
), `Gross Margin %` = c(44.56, 46.96, 13.48, 14.49), `Selling, General, & Admin. Expense` = c(419,
323.9, 424, 406), `Impairment Of Capital Assets` = c(0, 0,
2, 0), Advertising = c(1, 1, 1, 1), `Research & Development` = c(115.1,
88.8, 0, 0), `Restructuring And Mergern Acquisition` = c(0,
0, 47, 0), `Other Operating Expense` = c(-5.68434188608e-14,
1.13686837722e-13, 121, 8), `Operating Income` = c(420.2,
451.5, 429, 430), `Operating Margin %` = c(19.62, 24.54,
5.94, 7.38), `Interest Income` = c(0.9, 0.7, 24, 13), `Interest Expense` = c(-152.1,
-101.6, -84, -113), `Net Interest Income` = c(-151.2, -100.9,
-60, -100), `Other Income (Expense)` = c(-52.8, -9.3, -5,
-115), `Non Operating Income` = c(-52.8, -9.3, -5, -115),
`Other Income (Minority Interest)` = c(0, 0, -15, -13), `Gain on Sale of Security` = c(-1.3,
-0.8, 0, 7), `Write Off` = c(1, 1, 1, 1), `Pre-Tax Income` = c(216.2,
341.3, 364, 215), `Tax Provision` = c(298, -126.4, -51, 424
), `Tax Rate %` = c(-137.84, 37.03, 14.01, -197.21), `Net Income (Continuing Operations)` = c(514.2,
214.9, 315, 653), `Net Income (Discontinued Operations)` = c(0,
0, 0, 0), `Net Income` = c(514.2, 214.9, 300, 640), `Net Margin %` = c(24.01,
11.68, 4.15, 10.99), `Preferred Dividends` = c(0, 0, 31,
0), `EPS (Basic)` = c(2.83, 1.28, 1.82, 4.38), `EPS (Diluted)` = c(2.76,
1.27, 1.4, 4.36), `Shares Outstanding (Diluted Average)` = c(186.2,
168.8, 214.7, 146.8), `Depreciation, Depletion and Amortization` = c(252.5,
175.9, 277, 182), EBITDA = c(620.8, 618.8, 725, 510)), .Names = c("year",
"month", "company", "Revenue", "Cost of Goods Sold", "Gross Profit",
"Gross Margin %", "Selling, General, & Admin. Expense", "Impairment Of Capital Assets",
"Advertising", "Research & Development", "Restructuring And Mergern Acquisition",
"Other Operating Expense", "Operating Income", "Operating Margin %",
"Interest Income", "Interest Expense", "Net Interest Income",
"Other Income (Expense)", "Non Operating Income", "Other Income (Minority Interest)",
"Gain on Sale of Security", "Write Off", "Pre-Tax Income", "Tax Provision",
"Tax Rate %", "Net Income (Continuing Operations)", "Net Income (Discontinued Operations)",
"Net Income", "Net Margin %", "Preferred Dividends", "EPS (Basic)",
"EPS (Diluted)", "Shares Outstanding (Diluted Average)", "Depreciation, Depletion and Amortization",
"EBITDA"), row.names = c(NA, 4L), class = "data.frame")
Constants:
startDate <- "2012-01-01"
endDate <- "2016-12-31"
What I want: to create a function that applies a custom function to all numeric columns. I am trying to calculate CAGRs. The CAGR formula is as such:
((End Value / Beginning Value)^(1/number of years)-1)
So as you can see, I need for each column to be able to find the correct end value and beginning value.
My function right now is this:
cagr <- function(startval,endval,x,y,years){
return(((endval[x == year(endDate)]/startval[y == year(startDate)])^(1/(years-1)))-1)
}
cagrNew <- function(df,colum,x,y,years){
colum <- quo(colum)
x <- quo(x)
y <- quo(y)
out <- df %>%
group_by(!!company) %>%
summarise(xxxx = cagr(!!colum[!!x == year(endDate)],!!colum[!!y == year(startDate)],!!x,!!y,numYears))
return(out)
}
When I run the above function (cagrNEW), I get this error:
Error in `[.formula`(colum, !(!x == year(endDate))) :
attempt to set an attribute on NULL
My desired output:
Company RevenueCagr Cost of Goods Sold CAGR ....
ALSN .5% .3%
DAN .3% .2%

Answer 1

I haven't repeated the data above, to conserve space. Convert to tibble and assign.
# df <- as_tibble(...)
library(tidyverse)
library(scales) #< For percentage formatting
start_year <- 2012
end_year <- 2016
df %>%
filter(year %in% c(start_year, end_year)) %>%
group_by(company) %>%
arrange(desc(year), .by_group = TRUE) %>%
summarise_if(is.double, funs(CAGR = percent( (.[[1]]/.[[2]])^ (1/(end_year - start_year)) - 1) ) )
# CAGR = ((End Value / Beginning Value)^(1/number of years)-1)
#Checksum: ALSN Company, Revenue
# (End Value / Beginning Value)^((1/number of years))-1
percent(( (1840 / 2142) ^ (1/(2016-2012)) - 1))
#> [1] "-3.73%"