The :btree
operation creates a sorted index, it is similar to :hash
apart from the resulting nested-map is sorted instead of hash-ordered.
Like other index operations, this is considered a tuning parameter and isn't necessary for relic to work, typically you would only materialize these to enable faster query in advanced situations, or index access to sorted indexes.
:btree
indexes can enable range query optimisations, such as seqing from a key in a particular order.
Here is an example of indexing events by timestamp.
;; QUERY
;; you index expressions, here I am going to use a single expression,
;; but I could use multiple and get a nested map, e.g [:btree :a, :b, :c]
[[:from :Event]
[:btree :ts]]
;; STATE
{:Event [{:ts 0, :msg "hello"}
{:ts 1, :msg ", world"}
{:ts 2, :msg "!"}]}
;; INDEX
;; a sorted map, that supports subseq and rsubseq
;; get with rel/index
{0 #{{:ts 0, :msg "hello"}}
1 #{{:ts 1, :msg ", world"}}
2 #{{:ts 2, :msg "!"}}}
Can you improve this documentation?Edit on GitHub
cljdoc is a website building & hosting documentation for Clojure/Script libraries
× close