how does one simultaneously iterate on multiple fl...
# coroutines
j
how does one simultaneously iterate on multiple flows in the same loop to get the same index? without calling toList() first
e
zip
j
for 1000 flows?
e
Why not. But what’s the use-case?
j
im doing group by key..n, list<value...>
e
How you ended up with 1000 identically indexed flows?
j
mmap
fixed width files with .. obviosly fixed columns
e
Why would you need flows here? It is totally synchronous
j
the syncronous code works to n rows but im trying to switch out Array<Any?> with Flow<Any?> in a few places to remedy heap explosion in the intermediaries
e
Use sequence
j
and GroupBy is actually converting Array<T> to Array<Array<T>>
does flow->sequence->iterator violate any principals or is it just waiting to gel ? my impression is that sqeuences are going to make multiprocessing slightly more mental accounting than calling flatMapMerge
with the exception of zip i see no other multiple flow collector that acts like a latch syncronizer
the idea started as having n cores spaced at 2GB intervals to cleanly divide source decoders at the bottom of the draw-thru driver layer.