command complexity / Big-O

[mrkamel]

Hi and thanks for ardb!

It would be super-helpful to have some doc listing the complexity of the commands, optimally using Big-O notation, like redis does in its docs.

[yinqiwen]

the storage engine store data in disk, and there is cache for some operations, it's hard to estimate the complexity.
and since rocksdb/leveldb use tree like structure to store data, the basic search complexity would be log2(N).

[shailesh33]

+1 on the request. Essentially what I am looking for is, if I am doing a HSET, how many reads and writes does it translate to. Similarly for sorted set operations etc.

On Wed, Dec 13, 2017 at 7:05 PM, yinqiwen notifications@github.com wrote:

the storage engine store data in disk, and there is cache for some operations, it's hard to estimate the complexity. and since rocksdb/leveldb use tree like structure to store data, the basic search complexity would be log2(N).

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/yinqiwen/ardb/issues/394#issuecomment-351595121, or mute the thread https://github.com/notifications/unsubscribe-auth/ALle5-jAxoCFb-O9uywR_N8MvW9tygJLks5tAJBygaJpZM4Q7R8U .

[mrkamel]

yes, let's forget about the storage layer complexity. I was as well only talking about the complexity that is ardb specific, because even from looking into the ardb code it's not always easy to estimate it, even though it is well written - and btw, who wants to always look into the code to estimate the complexity of ZADD, HSET or SREM.

[shailesh33]

@yinqiwen Appreciate your response here. thanks.