FMS2: Safe inspection of unlimited dim name #386
Conversation
The FMS2 function `get_unlimited_dimension_name` raises a netCDF error if no unlimited dimension is found. This is problematic for legacy or externally created input files which may have not identifed their time axis as unlimited. This patch adds a new function, `find_unlimited_dimension_name` which mirrors the FMS2 function but returns an empty string if none are found. This is an internal function, not intended for use outside of the module.
Codecov Report
@@ Coverage Diff @@
## dev/gfdl #386 +/- ##
============================================
- Coverage 47.45% 38.22% -9.23%
============================================
Files 41 269 +228
Lines 4550 76359 +71809
Branches 798 14025 +13227
============================================
+ Hits 2159 29192 +27033
- Misses 2210 41928 +39718
- Partials 181 5239 +5058 see 254 files with indirect coverage changes 📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more |
|
Based on @jiandewang's comments, I think this is ready to merge (assuming no regressions). |
There was a problem hiding this comment.
This looks like a simple and elegant solution to the problems with using FMS2 to read a restart (or other) file without an unlimited time axis that are discussed at github.com//issues/380. Given the reports from EMC in the discussion of that issue that this commit addresses their problems, I am satisfied that these fairly simple changes have been adequately tested.
|
Gaea regression: https://gitlab.gfdl.noaa.gov/ogrp/MOM6/-/pipelines/19646 |
The FMS2 function
get_unlimited_dimension_nameraises a netCDF error if no unlimited dimension is found. This is problematic for legacy or externally created input files which may have not identifed their time axis as unlimited.This patch adds a new function,
find_unlimited_dimension_namewhich mirrors the FMS2 function but returns an empty string if none are found.This is an internal function, not intended for use outside of the module.