Make a adjacency matrix in R - r

Question

I want to make an adjacency matrix from a dataframe (mydata) consisting several rows with following rule:
List all letters as a square matrix
Count and sum number of connection from source from rest of columns (p1 p2 p3 p4 p5) of corresponding rows. For example, b is connected with a (2 and 8 rows) 5 times.
If letter is not included in source , connection values should be zero.
The dataframe is:
mydf <- data.frame(p1=c('a','a','a','b','g','b','c','c','d'),
p2=c('b','c','d','c','d','e','d','e','e'),
p3=c('a','a','c','c','d','d','d','a','a'),
p4=c('a','a','b','c','c','e','d','a','b'),
p5=c('a','b','c','d','I','b','b','c','z'),
source=c('a','b','c','d','e','e','a','b','d'))
The adjacency matrix should be as following
a b c d e g I z
a 4 2 1 3 0 0 0 0
b 5 1 3 0 1 0 0 0
c 1 1 2 1 0 0 0 0
d 1 2 3 2 1 0 0 1
e 0 2 1 3 2 1 1 0
g 0 0 0 0 0 0 0 0
I 0 0 0 0 0 0 0 0
z 0 0 0 0 0 0 0 0
I have hundreds of columns and thousands of rows. I would appreciate having any fastest way to do it in R

Answer 1

In base R, we can use table :
vals <- unlist(mydf[-ncol(mydf)])
table(factor(rep(mydf$source, ncol(mydf) - 1), levels = unique(vals)), vals)
# vals
# a b c d e g I z
# a 4 2 1 3 0 0 0 0
# b 5 1 3 0 1 0 0 0
# g 0 0 0 0 0 0 0 0
# c 1 1 2 1 0 0 0 0
# d 1 2 3 2 1 0 0 1
# e 0 2 1 3 2 1 1 0
# I 0 0 0 0 0 0 0 0
# z 0 0 0 0 0 0 0 0
In tidyverse we can do :
library(dplyr)
library(tidyr)
mydf %>%
pivot_longer(cols = -source) %>%
count(source, value) %>%
pivot_wider(names_from = value, values_from = n) %>%
complete(source = names(.)[-1]) %>%
mutate_all(~replace_na(., 0))