Reputation: 19375
I am using Dask 0.10
with the latest Anaconda
distribution and I run into the following error after calling .compute(get=dask.multiprocessing.get)
:
File "<ipython-input-8-cd6a1c9a02b6>", line 309, in find_matches
merged = data.merge(matches, on='url', how='left')
File "C:\Users\user\AppData\Local\Continuum\Anaconda2\lib\site-packages\dask\dataframe\core.py", line 1664, in merge
suffixes=suffixes, npartitions=npartitions)
File "C:\Users\user\AppData\Local\Continuum\Anaconda2\lib\site-packages\dask\dataframe\multi.py", line 428, in merge
how, npartitions, suffixes)
File "C:\Users\user\AppData\Local\Continuum\Anaconda2\lib\site-packages\dask\dataframe\multi.py", line 257, in hash_join
lhs2 = shuffle(lhs, left_on, npartitions)
File "C:\Users\user\AppData\Local\Continuum\Anaconda2\lib\site-packages\dask\dataframe\shuffle.py", line 160, in shuffle
import partd
File "C:\Users\user\AppData\Local\Continuum\Anaconda2\lib\site-packages\partd\__init__.py", line 16, in <module>
from .zmq import Client, Server
File "C:\Users\user\AppData\Local\Continuum\Anaconda2\lib\site-packages\partd\zmq.py", line 33, in <module>
log('Import zmq')
File "C:\Users\user\AppData\Local\Continuum\Anaconda2\lib\site-packages\partd\zmq.py", line 29, in log
with open('log', 'a') as f:
IOError: [Errno 13] Permission denied: 'log'
I have no idea what is causing this. Does Dask
needs to write on the root directory? I am working on a machine where I dont have admin rights and where there is a firewall.
Upvotes: 0
Views: 2461
Reputation: 57251
Does Dask needs to write on the root directory?
No, Dask operates entirely within userspace, as does Anaconda generally. In the traceback provided it looks like the error is coming from trying to open a file named "log"
in the local directory. Perhaps you're running Dask from a place where you don't have write permission?
Upvotes: 1