-
There are two things dask does differently when computing std and mean for timedelta columns.
For dataframe timedelta columns are dropped and for series different dtype is returned (float64 or numpy.…
-
If DDFS / network is being loaded excessively while demanding jobs are running, nodes may start dropping which causes lots of temporary failures in jobs.
There should be ways to balance QoS between D…
-
In SQL, it's common to work w/ large data and aggregate or filter it down to few enough rows that it could be merged into a single partition in memory.
Today you can achieve this with something lik…
-
Initiall reported here: https://github.com/lmorabit/lofar-vlbi/issues/70
After converting killMS solutions to fulljones h5parms with killMS2H5parm.py, merging these with H5parm_collector.py behaves…
-
Hi and thanks for providing this amazing tool.
I am currently running dada on a set of very deep sequenced samples. Around 3-6M 300bp NextSeq reads per sample remain after filtering.
What i'm cu…
-
This may be the expected behavior? I discovered this issue because I thought that setting the partition size manually `ddf.repartition(partition_size="100MB")` to try and balance these partitions woul…
-
Two kinds of problems here (the cause is probably the same):
```
class A(models.Model):
id = models.CharField(max_length=50, primary_key=True)
class B(A):
pass
```
```
Expected :…
-
Thank you for your excellent work! I am very interested in adopting ddf in other networks.
However, I fail to setup ddf in my device with error message "nvcc fatal : Unsupported gpu architecture '…
-
**What happened**:
Due to a mistake in our code, we were persisting a dask dataframe in one scheduler, but then ran the compute while specifying threads scheduler. What was weird was that the compu…
-
### ALL software version info
- MacOS - Sonoma 14.0
- python = 3.11.8
- notebook = 7.0.8
- dask = 2023.11.0
- datashader = 0.16.0
- geopandas = 0.14.2
- hvplot = 0.9.2
- holoviews = 1.18.3
-…