Why IN cluse is such peneltiy

SQL++
1

I’ve create this index:
CREATE INDEX channel_idx ON megatron (channel)

I have 100000 documents in the bucket

And i’m doing those queries:
1)
SELECT duration FROM megatron WHERE channel=1
~991ms

SELECT duration FROM megatron WHERE channel=1 OR channel=2
~1987ms

SELECT duration FROM megatron WHERE channel IN [1, 2]
~13028ms

Why the use of IN is such slow ?
are there any other option to optimize such query ? when I need multiple not ranged key ?

2

IN was improved in 4.5.1.

Can you use Couchbase 4.5.1 with your current index.

And then try the following index with 4.5.1.

CREATE INDEX idx_covering ON megatron(channel, duration);