fonty422
fonty422
I also had a method that was faster for large windows where rather than a windowed table, I had a standard table where the value was a list of times....
I also found [this class](https://github.com/faust-streaming/faust/blob/master/faust/windows.py) that seems to shwo a sliding window, and this [test example](https://github.com/robinhood/faust/blob/master/t/unit/windows/test_sliding_window.py) from the old robinhood version, but I have no idea how to implement it...
One possible solution that should be fast to aggregate across a large window would be to always increment the value in the current window, but if the current window is...
So in looking through the [test for sliding windows](https://github.com/faust-streaming/faust/blob/master/tests/unit/windows/test_sliding_window.py) I also happened to find the [table.py file](https://github.com/faust-streaming/faust/blob/master/faust/tables/table.py). Then by looking at how hopping and tumbling windows were implemented and cross...
A new solution that works very well, but still needs more smarts to work perfectly: - Events come into a stream processor (agent) and it increments a counter in the...
Thanks @doncat99. I did try using the on_window_close from the [example](https://github.com/robinhood/faust/issues/514) , but found that the closed/expired windows were 'lost' during a re-balance or after a crash of a worker...
Just in addition - @doncat99 correct me if I'm wrong, but using the hopping window won't be serviceable if we're looking to get 3 month aggregates at 1 minute intervals....
So after a little more looking, I managed to understand what you were asking me to do, @doncat99. I took a look at the [base.py](https://github.com/robinhood/faust/blob/master/faust/tables/base.py#L357-L379) file and put a few...
Just one additional point of information as to why we were hoping to not use a key to force it to be sent to the same partition every time. Of...
Thanks for the response @Roman1us. I'm only using memory until I can figure out how to install RocksDB - I'm very surprised just how painful a task that is an...