joseph nowak comments

Results 10 comments of


                                            joseph nowak

Is possible to modify a dataset once the serve is started?

I did a code that does exactly what I wanted, but It makes use of private attributes so I suppose It is not a good idea to use it and...

Call the persist method on an array after using map_blocks and sending a Python object never release the copies of the object created for every chunk

I think the issue is more related to dask distributed, so should I open the issue here https://github.com/dask/distributed/issues?

Call the persist method on an array after using map_blocks and sending a Python object never release the copies of the object created for every chunk

I modified the example to use the class MemorySampler. ```py3 import pandas as pd import dask import time from dask.distributed.diagnostics import MemorySampler ms = MemorySampler() class Mapper: def __init__(self, s:...

Call the persist method on an array after using map_blocks and sending a Python object never release the copies of the object created for every chunk

Is this a bug?

Call the persist method on an array after using map_blocks and sending a Python object never release the copies of the object created for every chunk

Thanks for the answer, it makes sense that it consumes that amount of memory due to the size of all the tasks but, is it expected that it never releases...

Call the persist method on an array after using map_blocks and sending a Python object never release the copies of the object created for every chunk

Thanks, I think that clarify all my doubts.

joseph nowak

Is possible to modify a dataset once the serve is started?

Call the persist method on an array after using map_blocks and sending a Python object never release the copies of the object created for every chunk

Call the persist method on an array after using map_blocks and sending a Python object never release the copies of the object created for every chunk

Call the persist method on an array after using map_blocks and sending a Python object never release the copies of the object created for every chunk

Call the persist method on an array after using map_blocks and sending a Python object never release the copies of the object created for every chunk

Call the persist method on an array after using map_blocks and sending a Python object never release the copies of the object created for every chunk

[Proposal] Add a Numba-optimized `rankdata` function

First execution of groupby on Xarray with Flox takes a lot of time

different `run_spec` between consecutive calls to `update_graph` | zarr-formatted xarray

different `run_spec` between consecutive calls to `update_graph` | zarr-formatted xarray