Composite key / secondary indexing strategy in Redis - data-structures

Question

Say I have some data of the following sort that I want to store in Redis:
* UUID
* State (e.g. PROCESSED, WAITING_FOR_RESPONSE)
* […] other vals
The UUID or the State are the only two variables that I will ever need to query on
What data structure in Redis is most suited for this?
How would I go about structuring the keys?

Answer 1

Okay, not sure I understand completely but going to try to go with it.
Assuming you need to look up all entities with state PROCESSED you can use sets for these:
SADD PROCESSED 123-abcd-4567-0000
Then you can easily find all entities with PROCESSED state. You'll do same for each state you want.
SMEMBERS PROCESSED
Now you'll also then want to have a hash for all your entities and its values:
HSET 123-abcd-4567-0000 state PROCESSED
HSET 123-abcd-4567-0000 otherproperty valuedata
This will set the "state" in the hash for the UUID to be processed (you'll need to figure out how to keep these in sync, you can use scripts or just handle in your code)
So in summary you have 2 major structures:
Sets to store your State to UUID information
Thus you have 1 set per state
Hashes to store the UUID to properties information
Thus you have 1 hash PER entity
Example Hash
123-abcd-4567-0000 => { state: PROCESSED, active: true }
987-zxy-1234-0000 => { state: PROCESSED, active: false }
But please clarify more if this doesn't seem to fit.
If you want to reduce your key space since the Hashes per entity can be much you can create a hash per attribute instead:
HSET states 123-abcd-4567-0000 PROCESSED
Thus you have a hash per attribute and your key is the UUID and value is the value of the property which is the hash key.
Example Hash
state => { 123-abcd-4567-0000: PROCESSED, 987-zxy-1234-0000: PROCESSED }
active => { 123-abcd-4567-0000: true, 987-zxy-1234-0000: false }

Answer 2

RediSearch (a Redis module) supports adding secondary index to existing data in Redis like Hash.
After setting the schema for the fiels you would like to index you can easily search based on these fields values.
e.g.
127.0.0.1:6379> FT.CREATE myIdx ON HASH PREFIX 1 doc: SCHEMA title TEXT
127.0.0.1:6379> hset doc:1 title "mytitle" body "lorem ipsum" url "http://redis.io"
(integer) 3
127.0.0.1:6379> FT.SEARCH myIdx "#title:mytitle" LIMIT 0 10
1) (integer) 1
2) "doc:1"
3) 1) "title"
2) "mytitle"
3) "body"
4) "lorem ipsum"
5) "url"
6) "http://redis.io"