Coord as cube #3485
Coord as cube #3485
Conversation
| daughter_coord = daughter_coords[0] | ||
|
|
||
| # Find a unique matching coord in the parent. | ||
| parent_coords = parent_cube.coord(daughter_coord) |
There was a problem hiding this comment.
this should be ".coords()" (plural) !
|
@pp-mo I'm not quite sure what the status of this PR is? Are you just sharing the concept? Or do you want it reviewed to be merged? Would you be able to clarify? |
Just sharing. It is really intended as more of a an "issue" : A proposal that we should provide a mechanism to convert such cube component to-and-from independent cubes. I wrote this stuff ages ago, but was reminded of it because it relates to the recent proliferation of "dimension metadata" components, in the effort to support Ancillary variables. So it would make good sense, if doing this now, to extend the idea to work with any of coords, ancillary variables and cell measures. |
|
Tried to change this to a newstyle non-mergeable PR, but it seems you don't get that choice. |
|
@pp-mo Would it make sense to close this PR and capture it as an issue? I'm all for having WIP PRs but it's confusing as a reviewer to determine what's a WIP PR that's actually aiming to deliver an agreed deliverable, what's a POC PR, and what's a "just sharing" for discussion PR. To me it makes sense to discuss as much as possible via an issue, then commit to doing the work in confidence that there is agreement and buy-in that the PR will be banked (hopefully) on review. What do you think? |
I think the problem is that I'm using a PR format to present proof-of-concept code. |
so
|
2 participants
WIP: incomplete, for discussion
Some ideas (incomplete) for extracting a coordinate as a cube.
And the reverse operation : putting a cube inside another cube (with matching dims) as an aux-coord.