Fix for 4-d variable output in standard memory mode #25
Conversation
| for key, val in field.items(): | ||
| if isinstance(val, str): | ||
| val = val.encode() |
There was a problem hiding this comment.
You'll need to undo all the changes to the string encoding that came in here. You can do that manually or by using something like source tree.
The reason this happened is that you copied in an old version of this file (that wasn't in git) into the current develop branch so you've just undone all the changes we've recently made to the file, that git has recorded. Does that make sense?
|
|
|
fixes #24 |
| yc = 0 | ||
| dd = np.zeros(self.f.variables[name][:,ys,xs].shape) | ||
| for p in points: | ||
| dd[:,yc] = p.df[name].values[self.slice] |
There was a problem hiding this comment.
This seems like we're just writing the same data to all of these points.
|
I'll have to look at this a bit more. I don't love either implementation so I'll need to think on it. I gave you some comments to get you started cleaning it up a bit though. Thanks. |
|
|
|
|
|
@anewman89 : Just ran across this - is this still an active PR? |
|
I never got anywhere further with this. In fact, my fix is likely buggy as I got some large disk I/O issues when I ran it on Yellowstone at NCAR. So lets close this specific request without a merge, I can make the original reason for the pull request an issue. |
|
Agreed - open an issue and then close this PR |
Issuing pull request for potential bug fix in 4-d variable output for
standardmemory mode.