Add zremrangebyrank

[peng-huang-ch]

Add zremrangebyrank to limit useful information count

[noamshemesh]

Thanks @penghap
On another thought, assuming we have an expiry for keys, don't you think the time of finding the indexes to delete isn't worth it?

[peng-huang-ch]

Yes, in the limit times, I found redis record lots of request time, but only the last few are useful. I think this can help us to reduce redis storage. Noam Shemesh <[email protected]> 于2019年3月2日周六 下午8:39写道：

…

Thanks @penghap <https://github.com/penghap> On another thought, assuming we have an expiry for keys, don't you think the time of finding the indexes to delete isn't worth it? — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#47 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AVBpLFXbpYrH92gZ_4PqYJSawSrqGdCIks5vSnD0gaJpZM4baW4c> .

[noamshemesh]

Please add a flag that is disabled by default for this. A possible name for this property: “tidy”

…

On Sat, 2 Mar 2019 at 14:52 hap ***@***.***> wrote: Yes, in the limit times, I found redis record lots of request time, but only the last few are useful. I think this can help us to reduce redis storage. Noam Shemesh ***@***.***> 于2019年3月2日周六 下午8:39写道： > Thanks @penghap <https://github.com/penghap> > On another thought, assuming we have an expiry for keys, don't you think > the time of finding the indexes to delete isn't worth it? > > — > You are receiving this because you were mentioned. > Reply to this email directly, view it on GitHub > <#47 (comment)>, > or mute the thread > < https://github.com/notifications/unsubscribe-auth/AVBpLFXbpYrH92gZ_4PqYJSawSrqGdCIks5vSnD0gaJpZM4baW4c > > . > — You are receiving this because you commented. Reply to this email directly, view it on GitHub <#47 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AE4fwWSqk9K6OhvsaAoZihxANlZozBmlks5vSnQRgaJpZM4baW4c> .

[peng-huang-ch]

Ok, I will do it, thanks. Noam Shemesh <[email protected]> 于2019年3月2日周六 下午8:57写道：

…

Please add a flag that is disabled by default for this. A possible name for this property: “tidy” On Sat, 2 Mar 2019 at 14:52 hap ***@***.***> wrote: > Yes, in the limit times, I found redis record lots of request time, but > only the last few are useful. I think this can help us to reduce redis > storage. > > Noam Shemesh ***@***.***> 于2019年3月2日周六 下午8:39写道： > > > Thanks @penghap <https://github.com/penghap> > > On another thought, assuming we have an expiry for keys, don't you think > > the time of finding the indexes to delete isn't worth it? > > > > — > > You are receiving this because you were mentioned. > > Reply to this email directly, view it on GitHub > > <#47 (comment) >, > > or mute the thread > > < > https://github.com/notifications/unsubscribe-auth/AVBpLFXbpYrH92gZ_4PqYJSawSrqGdCIks5vSnD0gaJpZM4baW4c > > > > . > > > > — > You are receiving this because you commented. > > > Reply to this email directly, view it on GitHub > <#47 (comment)>, > or mute the thread > < https://github.com/notifications/unsubscribe-auth/AE4fwWSqk9K6OhvsaAoZihxANlZozBmlks5vSnQRgaJpZM4baW4c > > . > — You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub <#47 (comment)>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AVBpLB2nhsdIPBfeSNoJoxjC6zSQ37_kks5vSnVBgaJpZM4baW4c> .

[peng-huang-ch]

@noamshemesh The tests have failed, seems to be the reason for the node version.

[noamshemesh]

That's fine, I'll drop support in node 4 and add the next versions. Thanks

[noamshemesh]

Merged. Thank you!

[Kikobeats]

Hello, I checked this PR in order to determinate the benefits to set tidy=true but I didn't find the point.

I understand that, since you are counting thing in redis, you don't want to count more than max, so you start deleting old counting over the same thing.

My question is: why not is this the default behavior? is it something related to perf? is it doesn't matter?

Just the thing I want to determinate is when I should use it 🙂

(after you reply I can do another PR for extend the explanation on docs)

[peng-huang-ch]

Yes, if we clean up redundant data every time, It's will loss the perf. But when a los of records for a key in the redis, It will improve it. I just provides a choice. When and where use it, just depend on the actual situation 😄

[Kikobeats]

Thanks for the comment!

Are you using it in a real production environment?

How many records do you consider would be enough to be significant enable it?

Also, thinking could be interesting a strategy that uses tidy after N times 🤔

[peng-huang-ch]

Yes, I have already used. But in the service i need to limit the count.
I think the max records is enough.
It sounds interesting, but it may need another way to record the N.