N1QL: Array Indexing: Index created without distinct is not being used in same way as created with distinct

Indexes created:
CREATE INDEX `idxjoin_yr` ON `default`(array `v` for `v` in `join_yr` end) WITH(

{"index_type" : "forestdb"}

)

and

CREATE INDEX `idxjoin_yr2` ON `default`((distinct (array `v` for `v` in `join_yr` end))) WITH(

{"index_type" : "forestdb"}

)

Query issued:
EXPLAIN select name from default use index (`idxjoin_yr2`) where any v in default.join_yr satisfies v = 2016 END;

uses the correct index

while same query with idxjoin_yr:
EXPLAIN select name from default use index (`idxjoin_yr`) where any v in default.join_yr satisfies v = 2016 END;

uses primary scan instead of index scan.

Dataset: Bucket default with documents like:
{
"tasks": [
{
"Marketing": [

{ "region2": "International", "region1": "South" }

,

{ "region2": "South" }

],
"Developer": [
"IOS",
"Indexing"
]
},
"Sales",
"QA"
],
"name": [

{ "FirstName": "employeefirstname-9" }

,

{ "MiddleName": "employeemiddlename-9" }

,

{ "LastName": "employeelastname-9" }

],
"address": [
[

{ "city": "Delhi" }

,

{ "street": "12th street" }

],
[

{ "country": "EUROPE", "apartment": 123 }

]
],
"email": "9-mail@couchbase.com",
"mutated": 0,
"hobbies": {
"hobby": [

{ "sports": [ "Badminton", "Football", "Basketball" ] }

,

{ "dance": [ "hip hop", "bollywood", "contemporary" ] }

,
"art"
]
},
"department": "Support",
"join_yr": [
2013,
2015,
2012
],
"_id": "query-testemployee10153.1877827-0",
"VMs": [

{ "RAM": 10, "os": "ubuntu", "name": "vm_10", "memory": 10 }

,

{ "RAM": 10, "os": "windows", "name": "vm_11", "memory": 10 }

,

{ "RAM": 10, "os": "centos", "name": "vm_12", "memory": 10 }

,

{ "RAM": 10, "os": "macos", "name": "vm_13", "memory": 10 }

]
}