Merge dataless #6741
Merge dataless #6741
Conversation
Codecov Report❌ Patch coverage is
Additional details and impacted files@@ Coverage Diff @@
## main #6741 +/- ##
==========================================
+ Coverage 90.29% 90.32% +0.02%
==========================================
Files 91 91
Lines 24656 24674 +18
Branches 4618 4623 +5
==========================================
+ Hits 22264 22286 +22
+ Misses 1620 1618 -2
+ Partials 772 770 -2 ☔ View full report in Codecov by Sentry. 🚀 New features to boost your workflow:
|
Status Update Fri 2023-10-10I'm currently using this branch to add some more extensive documentation, since I felt I needed a decent location to describe the properties of merging with dataless cubes. Also, I want to at least refer to #6739 Hence put this in draft for a while... |
| # We expect that the "parts" should **all be the same** | ||
| assert data.shape == part_shape | ||
| assert data.dtype == part_dtype |
There was a problem hiding this comment.
The assertion on shape is only made for dataful cubes; there is no check that the dataless cube shape is consistent. Should that be happening here too?
Can cubes of different shape end up in this function?
| else: | ||
| # We expect that the "parts" should **all be the same** | ||
| assert data.shape == part_shape | ||
| assert data.dtype == part_dtype |
There was a problem hiding this comment.
Raise exception instead?
Ruff S101: Favour exceptions over asserts
|
|
||
| Most notably, dataless cubes can be used as the target "grid cube" for most regridding | ||
| schemes, since in that case the cube's coordinates are all that the method uses. | ||
| See also :meth:`iris.util.make_gridcube`. |
There was a problem hiding this comment.
The implication seems to be that make_gridcube should be able to return dataless cubes, this doesn't look like part of this PR, is there a plan to implement this? If not, it might be worth clarifying what is relevant here about make_gridcube.
3 participants
Closes #6740
TBH this might have targetted "save_dataless" #6739 instead.
But now it can just wait till after that one