Is it not possible to use pipeline in get_many

[debugger22]

I am using get_many to retrieve multiple keys at the same time. Surprisingly unlike set_many, it calls the Redis individually for each key in the list.

Is there a way to optimize this behavior?

[WisdomPill]

it is using pipeline so it sends one command to the server.

Have a look here.

Do you think it is slow?

Probably using MSET is faster but it complicated things because of preprocessing of keys and versioning.

Do you fancy opening a PR for switching over to MSET

[tony]

Thank you

Have a look here.

That's just a link to redis docs - I find their docs a bit tricky to connect to real world scenarios 😆. When I opened this issue I was more curious about django-redis.

It would be great to see more: e.g. links to the current source / explanation of the codepath for django-redis' implementation of that.

[WisdomPill]

get_many (source here) uses MGET command which has a complexity of O(1) from django-redis perspective and a complexity of O(N), the complexity is all on redis side so if you are asking for many keys then it can be slow because of size of values and number of values.

While set_many (source here) just recently started using PIPELINE in order to improve total RTT.

Are you having issues reading many keys of writing?

To be honest I don't see a reason why it could be slow and a way to make it faster, but I would be happy to be proven wrong.