merging with changing the form of columns - r

Question

I have two data frames named "df" and "df1". what i want is merging df with df1 based on gender and district in such a way that after merging I only have one column of "prob.dis". more clearly, I want that if dis is 1, then the value from prob.dis1` should be used, and if dis is 5, then I want the value from prob.dis5.any help would be appreciated.
df<-
age gender dis
10 1 1
11 2 5
10 1 4
11 2 2
10 1 1
11 2 2
10 1 4
11 2 5
10 1 3
11 2 3
df1<-
age gender prob.dis1 prob.dis2 prob.dis3 prob.dis4 prob.dis5
10 1 0.0099 0.0124 0.0037 0.0176 0.1
11 2 0.0021 0.802 0.005 0.0029 0.2

Answer 1

Transform df1 into long format and join it with df on gender and dis:
library(dplyr)
library(tidyr)
df1 %>%
pivot_longer(
cols = matches("^prob.dis"),
names_to = c(".value", "dis"),
names_pattern = "([^0-9]+)([0-9]+)"
) %>%
mutate(dis = as.integer(dis)) %>%
select(-age) %>%
left_join(df, ., by = c("gender", "dis"))
age gender dis prob.dis
<dbl> <dbl> <dbl> <dbl>
1 10 1 1 0.0099
2 11 2 5 0.2
3 10 1 4 0.0176
4 11 2 2 0.802
5 10 1 1 0.0099
6 11 2 2 0.802
7 10 1 4 0.0176
8 11 2 5 0.2
9 10 1 3 0.0037
10 11 2 3 0.005
data:
df <- structure(list(age = c(10, 11, 10, 11, 10, 11, 10, 11, 10, 11
), gender = c(1, 2, 1, 2, 1, 2, 1, 2, 1, 2), dis = c(1, 5, 4,
2, 1, 2, 4, 5, 3, 3)), row.names = c(NA, -10L), class = c("tbl_df",
"tbl", "data.frame"))
df1 <- structure(list(age = c(10, 11), gender = c(1, 2), prob.dis1 = c(0.0099,
0.0021), prob.dis2 = c(0.0124, 0.802), prob.dis3 = c(0.0037,
0.005), prob.dis4 = c(0.0176, 0.0029), prob.dis5 = c(0.1, 0.2
)), row.names = c(NA, -2L), class = c("tbl_df", "tbl", "data.frame")
)

Answer 2

Update:
pivot_longer df2
right_join with df1
library(dplyr)
library(tidyr)
df1 <- df2 %>%
pivot_longer(
cols = starts_with("prob"),
names_to = "dis",
values_to = "prob.dis"
) %>%
mutate(dis = as.numeric(gsub("\\D", "", dis))) %>%
right_join(df1, by = c("age", "gender", "dis"))
Output:
age gender dis prob.dis
<int> <int> <dbl> <dbl>
1 10 1 1 0.0099
2 10 1 1 0.0099
3 10 1 3 0.0037
4 10 1 4 0.0176
5 10 1 4 0.0176
6 11 2 2 0.802
7 11 2 2 0.802
8 11 2 3 0.005
9 11 2 5 0.2
10 11 2 5 0.2

Answer 3

In case dis is only ranging from 1 to 5 and is sorted in the columns of df2 you can use match to find the row and use 2 + df$dis to get the column, which can be subseted with a matrix produced with cbind.
df$prop.dis <- df1[cbind(match(df$gender, df1$gender), 2 + df$dis)]
df
# age gender dis prop.dis
#1 10 1 1 0.0099
#2 11 2 5 0.2000
#3 10 1 4 0.0176
#4 11 2 2 0.8020
#5 10 1 1 0.0099
#6 11 2 2 0.8020
#7 10 1 4 0.0176
#8 11 2 5 0.2000
#9 10 1 3 0.0037
#10 11 2 3 0.0050
or using the names of df1 to match the colums:
df$prop.dis <- df1[cbind(match(df$gender, df1$gender)
, match(paste0("prob.dis", df$dis), names(df1)))]
In case also age should be matched use in addition interaction:
M <- c("age", "gender")
df$prop.dis <- df1[cbind(match(interaction(df[M]), interaction(df1[M])), 2 + df$dis)]