Using N1QL reserved words - couchbase

Question

I have a document with the structure { doc.data.role.level }, I want to access to level property,
but using this query
SELECT r.data.role.level FROM `hostel` r where r.id = '12345678Z'
I got the error
"msg": "syntax error - at role",
but with those query I got no results
SELECT r.data.`role`.level FROM `hostel` r where r.id = '12345678Z'
SELECT `r.data.role.level` FROM `hostel` r where r.id = '12345678Z'

Answer 1

If there are special characters in fields/identifiers or reserve keywords you must escape them with back-ticks(i.e. escaped identifiers) as described https://docs.couchbase.com/server/current/n1ql/n1ql-language-reference/identifiers.html
Example 1:
"role" is reserve keyword
{
"id":"12345678Z",
"data":{"role":{"level":3}}
}
SELECT r.data.`role`.level
FROM hostel AS r
WHERE r.id = "12345678Z";
Example 2:
Field has special character dot. You must escape whole field name (to interpret as field vs nested document).
{
"id":"12345678Z",
"data.role.level":5
}
SELECT r.`data.role.level`
FROM hostel AS r
WHERE r.id = "12345678Z";
Example 3:
hostels is array. If you are looking inside array you must use ANY syntax.
{
"hostels": [ { "id":"12345678Z"},
{ "id":"34545678Z"}}
]
"data":{"role":{"level":3}}
}
SELECT r.data.`role`.level
FROM hostel AS r
WHERE ANY h IN r.hostels SATISFIES h.id = "12345678Z" END;

Answer 2

You're on the right track with backticks to escape reserved words.
If you aren't getting an error message, then I suspect it's your WHERE clause. Are you sure your documents have an "id" field in them?
{ "id": "12345678Z", "foo": "bar", ... etc ... }
Or are you trying to query based on the document ID (in the below screenshot there is no "id" field in the document)?
If the latter, then you should use META().id:
SELECT * FROM `testbucket` r where META(r).id = '12345678Z'
I know this might seem a bit confusing, but "id" isn't a magic field in Couchbase. It's a field just like any other. The actual ID of the document is in the document's metadata.
Also, a recommendation: if you are trying to get a single document by its ID, you're much better off using the key/value API instead of a N1QL query. No need to involve the overhead of query/index services.