Multiply columns from different data frames if dates match - r

Question

I have the following two data frames:
df1 <- data.frame(Category = c("A", "A", "A", "B", "B", "B", "C", "C", "C"),
Date = c(2001, 2002, 2003, 2001, 2002, 2003, 2001, 2002, 2003),
Beta1 = c(1, 3, 4, 4, 5, 3, 5, 3, 1),
Beta2 = c(2, 4, 6, 1, 1, 2, 5, 4, 2))
df2 <- data.frame(Date = c(2001, 2002, 2003),
Column1 = c(10, 20, 30),
Column2 = c(40, 50, 60))
Say I assign category A to Column1 and and category C to Column2. I want to multiply the row value from Column1 with the row betas from category A, if the dates match. Similarly, I want to multiply the row value from Column2 with the row betas from category C, if the dates match.
The match between a category and a column is of my own choosing. Assigning this myself won’t be a problem I think because I have relatively few columns.
Preferably, I want the output to look like this:
results <- data.frame(Date = c(2001, 2002, 2003),
Column1_categoryA_beta1 = c(10, 60, 120),
Column1_categoryA_beta2 = c(20, 80, 180),
Column2_categoryC_beta1 = c(200, 150, 60),
Column2_categoryC_beta2 = c(200, 200, 120))
Any help in how I best can approach this problem is very much appreciated!

Answer 1

With some data wrangling using tidyr and dplyr this can be achieved like so:
df1 <- data.frame(Category = c("A", "A", "A", "B", "B", "B", "C", "C", "C"),
Date = c(2001, 2002, 2003, 2001, 2002, 2003, 2001, 2002, 2003),
Beta1 = c(1, 3, 4, 4, 5, 3, 5, 3, 1),
Beta2 = c(2, 4, 6, 1, 1, 2, 5, 4, 2))
df2 <- data.frame(Date = c(2001, 2002, 2003),
Column1 = c(10, 20, 30),
Column2 = c(40, 50, 60))
library(dplyr)
library(tidyr)
df2_long <- df2 %>%
pivot_longer(-Date, names_to = "Column", values_to = "Value") %>%
mutate(Category = ifelse(Column == "Column1", "A", "C"))
df2_long %>%
left_join(df1) %>%
mutate(Beta1 = Value * Beta1,
Beta2 = Value * Beta2) %>%
select(Date, Category, Column, Beta1, Beta2) %>%
pivot_wider(id_cols = Date, names_from = c("Column", "Category"), values_from = c("Beta1", "Beta2"))
#> Joining, by = c("Date", "Category")
#> Warning: Column `Category` joining character vector and factor, coercing into
#> character vector
#> # A tibble: 3 x 5
#> Date Beta1_Column1_A Beta1_Column2_C Beta2_Column1_A Beta2_Column2_C
#> <dbl> <dbl> <dbl> <dbl> <dbl>
#> 1 2001 10 200 20 200
#> 2 2002 60 150 80 200
#> 3 2003 120 60 180 120
Created on 2020-04-14 by the reprex package (v0.3.0)

Answer 2

One way to get there while keeping the Category variable in the final data frame is the following:
df3 <- left_join(df1, df2, by="Date")
df4 <- df3 %>%
group_by(Date, Category) %>%
mutate(Col1Bet1 = Column1 * Beta1, Col1Bet2 = Column1 * Beta2, Col2Bet1 = Column2 * Beta1, Col2Bet2 = Column2 * Beta2)
which gives the following:
# A tibble: 9 x 10
# Groups: Date, Category [9]
Category Date Beta1 Beta2 Column1 Column2 Col1Bet1 Col1Bet2 Col2Bet1 Col2Bet2
<fct> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl> <dbl>
1 A 2001 1 2 10 40 10 20 40 80
2 A 2002 3 4 20 50 60 80 150 200
3 A 2003 4 6 30 60 120 180 240 360
4 B 2001 4 1 10 40 40 10 160 40
5 B 2002 5 1 20 50 100 20 250 50
6 B 2003 3 2 30 60 90 60 180 120
7 C 2001 5 5 10 40 50 50 200 200
8 C 2002 3 4 20 50 60 80 150 200
9 C 2003 1 2 30 60 30 60 60 120

Answer 3

This could be a start. The result data.table has all information you want just in another format.
df3 <- merge(df1, df2)
df3$b1 <- ifelse(df3$Category=="A", df3$Beta1*df3$Column1, ifelse(df3$Category=="C", df3$Beta1*df3$Column2, NA))
df3$b2 <- ifelse(df3$Category=="A", df3$Beta2*df3$Column1, ifelse(df3$Category=="C", df3$Beta2*df3$Column2, NA))
# Date Category Beta1 Beta2 Column1 Column2 b1 b2
# 1 2001 A 1 2 10 40 10 20
# 2 2001 C 5 5 10 40 200 200
# 3 2001 B 4 1 10 40 NA NA
# 4 2002 A 3 4 20 50 60 80
# 5 2002 B 5 1 20 50 NA NA
# 6 2002 C 3 4 20 50 150 200
# 7 2003 B 3 2 30 60 NA NA
# 8 2003 A 4 6 30 60 120 180
# 9 2003 C 1 2 30 60 60 120