Dependents and Precedents in R - r

Question

Need help in flagging number of Dependents and Precedents in R. My data frame contains some formulas (strings) and I want to add "col3" which should contain: 0 for A1, 1 for A2 (Because A2 is dependent on A1 - One dependency) and 2 for A3 (Because A3 is dependent on A2/A1).
col1 <- c('A1','A2','A3', 'A6','A4','A7')
col2 <- c('X1+Y1','A1+Y2', 'A4+Y3+A2', 'Y5+A1','A2+A1+A3','A2+A1')
df <- data.frame(col1, col2, stringsAsFactors=F)
My Output should look like:
col1 col2 col3
1 A1 Y1 0
2 A2 A1+Y2 1
3 A3 A4+Y3+A2 5
4 A6 Y5+A1 1
5 A4 A2+A1+Y3 3
6 A7 A2+A1 3
I have a data frame with 100+ rows of this format. Appreciate if you could help with this.

Answer 1

Below code produces the correct output.
col0 <- c('A1','A2','A3', 'A6','A4','A7')
col2 <- c('X1+Y1','A1+Y2', 'A1+Y3+A2', 'Y5+A2','A2+A1+A3','A2+A3')
df <- data.frame(col0, col2, stringsAsFactors=F)
library(tidyr)
library(dplyr)
df1 <- df %>%
separate(col2, into = as.character(c(1:4)),sep = "\\+") %>%
replace(is.na(.),"")
df1$OOE <- 0
for (i in 1:nrow(df1)) {
for (j in 2:ncol(df1)) {
for (k in 1:nrow(df1)) {
if (df1[i,j] == df1$col0[k]) df1$OOE[i]=df1$OOE[k]+df1$OOE[i]+1
}
}
}
col0 1 2 3 4 OOE
1 A1 X1 Y1 0
2 A2 A1 Y2 1
3 A3 A1 Y3 A2 3
4 A6 Y5 A2 2
5 A4 A2 A1 A3 7
6 A7 A2 A3 6

Answer 2

If AX can have a dependency on AY where Y>X, we need a tree like structure to find the dependencies. I knew about the igraph package but it seems to complex for the task. We just need some reference semantics and after some research, data.tree package seems appropriate. Here is the code:
col1 <- c('A1','A2','A3', 'A6','A4','A7')
col2 <- c('X1+Y1','A1+Y2', 'A1+Y3+A2', 'Y5+A2','A2+A1+A3','A2+A3')
df <- data.frame(col1, col2, stringsAsFactors=F)
require(data.tree)
# Create the graph/forest based on the data
getForest <- function(data) {
res <- new.env()
for( i in 1:nrow(data)){
nname <- data$col1[i]
if(!exists(nname,where=res))
assign(nname,Node$new(nname), pos=res)
par <- get(nname, envir=res)
print(par)
#Add the childs
deps <- unlist(regmatches(data$col2[i],gregexpr("A\\d+",data$col2[i])))
for( ch in deps) {
print("Ammm")
if(!exists(ch, where=res))
assign(ch,Node$new(ch), pos=res)
child <- get(ch, envir=res)
par$AddChildNode(child)
}
}
#Return the nodes
res
}
f <- getForest(df)
# Function to get the dependency level
getLevel<- function(node) {
if (node$count == 0)
return (0)
else {
return (length(node$children)+sum(sapply(node$children,getlevel)))
}
}
#Add dependency level to data frame
df$col3 <- sapply(df$col1, function(x) {getLevel(get(x,f))})
df
# col1 col2 col3
#1 A1 X1+Y1 0
#2 A2 A1+Y2 1
#3 A3 A1+Y3+A2 3
#4 A6 Y5+A2 2
#5 A4 A2+A1+A3 7
#6 A7 A2+A3 6