how to separate a column value into multiple columns, in rule mining

Asked May 10 '18 at 07:41

Active May 10 '18 at 07:41

Viewed 32 times

I have below data frame where I want to separate my rule column into separate columns.

Data :

          rules    support  confidence  lift count
5   {32,39,6} => {65}   0.00055 1.00    22.3    12
6   {39,6,65} => {32}   0.00055 0.86    23.4    12
9   {14,16,26} => {4}   0.00051 0.79    8.9     11
13  {16,20,4} => {26}   0.00069 0.71    8.0     15
16  {16,26,33} => {4}   0.00055 0.67    7.6     12
1   {53} => {31}        0.00087 0.66    89.6    19

Expected output :

lhs          rhs 
{32,39,6}   {65}
{39,6,65}   {32}
{14,16,26}  {4}

I have tried the below code

rules = sapply(assoc_rules$rules,function(x){
    x = gsub("[\\{\\}]", "", regmatches(x, gregexpr("\\{.*\\}", x))[[1]])
    x = gsub("=>",",",x)
    x = str_replace_all(x," ","")
    return( x )
})

asked May 10 '18 at 07:41

pylearner

1,358
2
10
26

1

Why not use `strsplit`? `strsplit(assoc_rules$rules, " => ", fixed=TRUE)`? Or `df %>% separate(rules, c("lhs", "rhs"), " => ")` with `tidyr`? – Wiktor Stribiżew May 10 '18 at 07:45
1

@WiktorStribiżew Error in strsplit(assoc_rules$rules, " => ", fixed = TRUE): non-character argument Traceback: 1. strsplit(assoc_rules$rules, " => ", fixed = TRUE) – pylearner May 10 '18 at 07:59
Yes, just tried with a data frame, it worked with a vector. I did not get any errors using `separate` though. – Wiktor Stribiżew May 10 '18 at 08:02
1

@pylearner Maybe your `rules` column is a factor, convert it to character first. – Ronak Shah May 10 '18 at 08:05
@WiktorStribiżew Thanks Wiktor, I could now use the above mention code. – pylearner May 10 '18 at 09:03
Yeah, [it seems to work](https://ideone.com/hm6e3C). – Wiktor Stribiżew May 10 '18 at 09:21

how to separate a column value into multiple columns, in rule mining

0 Answers0