ERROR: is.subset: Error in .local(x, ...) : All item labels in x must be contained in 'itemLabels' or 'match'. about arules

[mhahsler]

Hi sjain777,

The following is an item encoding error.
Error in .local(x, ...) :
All item labels in x must be contained in 'itemLabels' or 'match'._

You get it because the two transaction sets that you create do not have the same encoding for items. You can see that the sets have a different number of items.

t1 <- as(list(c("salt","water"),c("pepper")), "transactions")
t1
transactions in sparse format with
2 transactions (rows) and
3 items (columns)

t2 <- as(list(c("salt","water")), "transactions")
t2
transactions in sparse format with
1 transactions (rows) and
2 items (columns)

I think your second example is caused by the same problem...

-Michael

So to make them comparable you need to do this

t2_new <- recode(t2, match = t1)
t2_new
transactions in sparse format with
1 transactions (rows) and
3 items (columns)

is.subset tries to do this internally, but fails for your data (I will fix that).
You can read up on item encoding in ?encode. The man page is somewhat hidden, but I will add some references in the other manual pages to make this easier to find. Now subset works as expected.

is.subset(t1, t2_new)
{salt,water}
{salt,water} TRUE
{pepper} FALSE

[sjain777]

Thanks for your quick response. I tried adding "recode" in my process as you advised above, but that failed too. I think, probably new labels (columns) in my test data are causing the problem this time. Here is the example to demonstrate the error:

t1 <- as(list(c("salt","water"),c("pepper"), c("honey"), c("lemon")), "transactions")
t1
transactions in sparse format with
4 transactions (rows) and
5 items (columns)

t2 <- as(list(c("salt","sugar", "water")), "transactions")
t2
transactions in sparse format with
1 transactions (rows) and
3 items (columns)

t2_new <- recode(t2, match = t1)
Error in .local(x, ...) :
All item labels in x must be contained in 'itemLabels' or 'match'.

Thanks for offering to work on a fix

• Supriya

[mhahsler]

You have in each set an item that is not in the other:

t1 <- as(list(c("salt","water"),c("pepper"), c("honey"), c("lemon")), "transactions")
t1
transactions in sparse format with
4 transactions (rows) and
5 items (columns)
t2 <- as(list(c("salt","sugar", "water")), "transactions")
t2
transactions in sparse format with
1 transactions (rows) and
3 items (columns)
itemLabels(t1)
[1] "honey" "lemon" "pepper" "salt" "water"
itemLabels(t2)
[1] "salt" "sugar" "water"

So you need to recode them using all items (union)

il <- union(itemLabels(t1), itemLabels(t2))
t1 <- recode(t1, il)
t2 <- recode(t2, il)

now subsetting will work

is.subset(t1,t2)
{salt,water,sugar}
{salt,water} TRUE
{pepper} FALSE
{honey} FALSE
{lemon} FALSE

[sjain777]

Thanks again for your quick feedback. Doing the "union" and then "using "recode" as you suggested above worked! Thanks!

[dr-habeba]

i try to make classification using cba algorithm in rpy2 but i found this error:

Traceback (most recent call last):

File "", line 1, in runfile('D:/dm.py', wdir='D:')

File "C:\ProgramData\Anaconda2\lib\site-packages\spyder\utils\site\sitecustomize.py", line 880, in runfile execfile(filename, namespace)

File "C:\ProgramData\Anaconda2\lib\site-packages\spyder\utils\site\sitecustomize.py", line 87, in execfile exec(compile(scripttext, filename, 'exec'), glob, loc)

File "D:/dm.py", line 104, in c_p=class_pred.pred_class_on_test(c, dtest)

File "C:\ProgramData\Anaconda2\lib\site-packages\rpy2\robjects\functions.py", line 178, in call return super(SignatureTranslatedFunction, self).call(*args, **kwargs)

File "C:\ProgramData\Anaconda2\lib\site-packages\rpy2\robjects\functions.py", line 106, in call res = super(Function, self).call(*new_args, **new_kwargs)

RRuntimeError: Error in .local(x, ...) : All item labels in x must be contained in 'itemLabels' or 'match'.

code:

import os
os.environ['R_HOME'] = 'C:/Program Files/R/R-3.5.1'
os.environ['R_USER'] = 'C:/ProgramData/Anaconda2/Lib/site-packages/rpy2'

import rpy2.robjects as rr
#to import any r backage
from rpy2.robjects.packages import importr

to call any r user defiend function from pythn string

from rpy2.robjects.packages import SignatureTranslatedAnonymousPackage as stap

to convert to / from python/r dataframe

from rpy2.robjects import pandas2ri
pandas2ri.activate()

base = importr('base')

read data set python

from sklearn.cross_validation import train_test_split
import pandas as pd
data= pd.read_csv("C://Users//El-Amir Tech//Desktop//tumor.csv")

print(data.info())

data=data.astype(str)

train,test=train_test_split(data, test_size=0.2, random_state=20)

to convert python data frame to r data frame

dd=pandas2ri.py2ri(train)
dtest=pandas2ri.py2ri(test)

data minin association rule function from r

rs=importr('arules')
clas=importr('arulesCBA')
rstring="""

generate rules

ruless <- function (d){

da<- as.data.frame(d)

trans <- as(da, "transactions")

to get rhs=outcome only

incom <- grep("^Class=",itemLabels(trans), value=TRUE)

rules <- apriori(trans, parameter = list(supp = 0.01, conf = 0.8,minlen=2, target = "rules"),appearance=list(rhs=incom))
return (rules)
}

convert rules to data frame

rule2dataframe<-function(r){
n<-DATAFRAME(r,separate=TRUE)
return (n)
}

apply cba classifier

classf <- function (rs){
c1<- CBA_ruleset(Class~ ., rs)
return (c1)
}

pred_class_on_test <- function(classifir,dtest) {
d<- as.data.frame(dtest)
results <- predict(classifir, d)

return(results)}

"""

call r function using stap

rule= stap(rstring,"ruless")
m=rule.ruless(dd)

r2d=stap(rstring,"rule2dataframe")
r=r2d.rule2dataframe(m)

convert rules r dataframe to pandas dataframe

rules=pandas2ri.ri2py_dataframe(r)
print (rules)

classs=stap(rstring,"classf")
c=classs.classf(m)
print c

class_pred=stap(rstring,"pred_class_on_test")

c_p=class_pred.pred_class_on_test(c, dtest)
print c_p

so how i can fix this error?

data set link :
https://archive.ics.uci.edu/ml/datasets/primary+tumor