https://pinot.apache.org/ logo
Join Slack
Powered by
# custom-aggregators
  • k
    you can work on the optimization in the next iteration, establish a baseline
  • d
    yes you are right
  • d
    i think basically i should look at the strcture like a tree with leafs
  • k
    yeah, you can look at it that way
  • k
    its just an encoding algorithm
  • d
    so for each point i need to understand if i need to extend or not a range...or just adjust the max or the min value onlnie
  • d
    ok
  • d
    sounds interesting
  • d
    🙂
  • k
    you dont need to maintain min/max in this case
  • k
    this is a simple encoding/compression problem
  • k
    think about how would compress a list of sorted integers
  • k
    1,2,3,4,5,10, 14,15,16 ...
  • k
    [1,5],10,[14-16]
  • k
    talk tomorrow
  • k
    enjoy 🙂
  • d
    yeah got it
  • d
    see you tomorrow!
  • d
    thanks
  • d
    the "timestamps" factor is imoprtant so i need to hold it somehow too...however..ok got it
  • d
    hello
  • d
    i did the code with python...just a rough code but it works as expected.. i surelly need to optimize it and then translate it to java
  • d
    i share the code, few minutes
  • d
  • d
    i shuffle the docs list and then see if the output is the same..
  • d
    sure...it is rough without comments too 🙂 but ... however i now can handle range from completely random data and without inner loops, just one.
  • d
    so for now i try to 1. clean the code 2. optimize this code and then i will go with java.
  • k
    @User has left the channel
  • b
    @User has left the channel
  • j
    @Juraj Pohanka has left the channel