Jaccard Similarity Check function doesn't work correctly

The following jaccard similarity check query finds records that have the same 
unordered list fields by giving threshold 1.0f, but the returned results are 
not correct. 

------------------------------
drop dataverse test if exists;
create dataverse test;
use dataverse test;

create type TwitterUserType as closed {
    screen-name: string,
    lang: string,
    friends-count: int32,
    statuses-count: int32,
    name: string,
    followers-count: int32
} 

create type TweetMessageType as closed {
    tweetid: int64,
        user: TwitterUserType,
        sender-location: point,
    send-time: datetime,
        referred-topics: {{ string }},
        interests: [string],
    message-text: string,
    countA: int32,
    countB: int32
}

create dataset TweetMessages(TweetMessageType)
primary key tweetid;

use dataverse test; 
for $t1 in dataset('TweetMessages')
for $t2 in dataset('TweetMessages')
let $sim := similarity-jaccard-check($t1.referred-topics, $t2.referred-topics, 
1.0f) 
where $sim[0] and $t1.tweetid != $t2.tweetid and $t1.tweetid < int64("10")
return { "msg1":$t1.message-text, "msg2":$t2.message-text}  

result:
{ "msg1": " dislike samsung the speed is OMG", "msg2": " hate samsung the speed 
is horrible:(" }
{ "msg1": " love sprint its voice-clarity is amazing", "msg2": " can't stand 
sprint the voice-clarity is bad" }
{ "msg1": " hate samsung the 3G is horrible", "msg2": " like samsung the 3G is 
mind-blowing:)" }
{ "msg1": " hate motorola its 3G is OMG:(", "msg2": " can't stand motorola the 
3G is OMG:(" }
{ "msg1": " hate sprint the wireless is terrible:(", "msg2": " like sprint its 
wireless is mind-blowing:)" }
{ "msg1": " love t-mobile the speed is mind-blowing:)", "msg2": " love t-mobile 
the speed is mind-blowing:)" }
{ "msg1": " hate iphone the shortcut-menu is bad", "msg2": " love iphone the 
shortcut-menu is mind-blowing" }
{ "msg1": " dislike samsung the speed is OMG", "msg2": " hate samsung its speed 
is horrible" }
{ "msg1": " like motorola its voice-command is amazing", "msg2": " hate 
motorola its voice-command is OMG" }
{ "msg1": " love sprint its voice-clarity is amazing", "msg2": " like sprint 
the voice-clarity is good:)" }
{ "msg1": " hate samsung the 3G is horrible", "msg2": " dislike samsung its 3G 
is bad:(" }
{ "msg1": " dislike samsung the speed is OMG", "msg2": " like samsung the speed 
is mind-blowing" }
{ "msg1": " love sprint its voice-clarity is amazing", "msg2": " can't stand 
sprint the voice-clarity is terrible" }
{ "msg1": " hate iphone the shortcut-menu is bad", "msg2": " like iphone the 
shortcut-menu is mind-blowing" }
{ "msg1": " hate iphone the shortcut-menu is bad", "msg2": " like iphone its 
shortcut-menu is amazing:)" }
{ "msg1": " hate verizon the 3G is OMG:(", "msg2": " love verizon its 3G is 
mind-blowing:)" }
{ "msg1": " hate samsung the 3G is horrible", "msg2": " dislike samsung its 3G 
is bad" }
{ "msg1": " hate motorola its 3G is OMG:(", "msg2": " dislike motorola its 3G 
is horrible:(" }
{ "msg1": " hate sprint the wireless is terrible:(", "msg2": " dislike sprint 
the wireless is OMG:(" }
{ "msg1": " love t-mobile the speed is mind-blowing:)", "msg2": " love t-mobile 
its speed is good:)" }
{ "msg1": " hate iphone the shortcut-menu is bad", "msg2": " like iphone its 
shortcut-menu is good" }
{ "msg1": " hate motorola its 3G is OMG:(", "msg2": " hate motorola the 3G is 
OMG:(" }
{ "msg1": " hate motorola its 3G is OMG:(", "msg2": " like motorola its 3G is 
good:)" }
{ "msg1": " dislike samsung the speed is OMG", "msg2": " can't stand samsung 
the speed is terrible:(" }
Original issue reported on code.google.com by kiss...@gmail.com on 18 May 2014 at 2:58
Attachments:
tw_for_indexleftouterjoin.adm
lwhay / asterixdb

Jaccard Similarity Check function doesn't work correctly #774
