Error “can coerce list with atomic components only” when transforming into transactions arules

Question

I have a data frame and it looks like this:

str(df)

'data.frame':   174671 obs. of  3 variables:
 $ COD_PRODUCT: Factor w/ 338 levels "001CH","002CO",..: 1 1 1 1 1 1 1 1 1 1 ...
 $ STORE      : Factor w/ 15 levels "Barcelona","Florencia",..: 4 13 4 5 11 12 5 13 1 12 ...
 $ SELL_ID    : Factor w/ 74327 levels "BA0000000","BA0000001",..: 28696 65976 27147 14291 51141 59023 12249 61636 5495 59314 ...

There is neither NAs nor invalid values in these Factors, as we can see here:

print(unlist(lapply(df, function(x) any(is.na(x)))))
COD_PRODUCT       STORE     SELL_ID 
      FALSE       FALSE       FALSE

My goal is to create a data mining model (association rules) and for that, I had previously successful imported the package arules. The next step that I followed was to split my df into a list and apply an operation to eliminate duplicated occurrences, such as:

ppvt <- split(x = df[, c("COD_PRODUCT", "STORE")], f = df$SELL_ID)
ppvt <- lapply (ppvt, unique)

So far, everything works fine. However, when I tried to transform into transactions:

ppvt <- as (ppvt,  "transactions" )

I'm getting an error message:

 Error in asMethod(object) : can coerce list with atomic components only

Totally clueless about it (and yes, as you can imagine, I searched a lot on the Internet before coming here). Do you have an idea about how to solve it?

Any help will be really appreciated.

Regards,

Can you provide a reproducible example of the issue (i.e., with code and data that we can download and run)? — Lyngbakr
Look at str(ppvt), and you will see that each element of the list is a dataframe. In order to convert to transactions you need each element of the list to be a vector. — Esther
Hi @Esther I think you're right. If you want to provide the answer below I will mark as correct giving you the credits :) Thanks for collaborating. — Fernando Barbeiro

nashank79 nashank79 · Accepted Answer · 2019-04-30T22:32:52

Can you check the class(ppvt) before your:

ppvt <- as (ppvt, "transactions" ).

It should return a data.frame for your:

ppvt <- as (ppvt, "transactions" )

to work.

I was working with dplyr and faced a similar problem. Converting the tibble to data.frame using as.data.frame() solved the problem.

Error “can coerce list with atomic components only” when transforming into transactions arules

1 Answers