Martin Yeo

Have investigated this afternoon. >it might be worth seeing what happens with a `(2, 2, 10000)` benchmark - is that still 1.5 times slower? This turned out to be true....

Just been chatting to @hsteptoe offline. He works with some software that insists on GeoPandas format, I think there are enough other geo-referenced tabular formats - mostly relying on polygon...

>I think my instinct is to add an `iris.geopandas` module I could see the case for not having the existing routines 'magically' do two different things, but I'd still rather...

> OK, so `from iris.pandas import as_geo_data_frame` would be a reasonable compromise? Yes that's the kind of thing I meant

> * [ ] Need to work out why there seems to be a dependency conflict with `geopandas` and `pyvista`, `vtk` and `geovista`. https://github.com/SciTools/iris/issues/5517#issuecomment-1771315944

>Perhaps in the case of VLEN datatypes, we should err on the side of caution and always load lazily? This should be fine. The code you reference was added in...

We've had reports that the `ncks` workaround doesn't work 100% of the time. The team are deployed on other priorities at the moment ([Backlog · 🐔Iris v3.11 (github.com)](https://github.com/orgs/SciTools/projects/37)), but here...

Hi @mo-cbradsha, I'm afraid I can't replicate your error. This still works for the original file @momichaelkendon sent me, and just to be sure I registered for a CDS account...

>given users already complain about warning fatigue with iris. [Filtering Warnings — Iris 3.11.1 documentation](https://scitools-iris.readthedocs.io/en/stable/further_topics/filtering_warnings.html)

Agreed solution between @pp-mo, @ukmo-ccbunney, @trexfeathers: align the handling of variable length strings with all our variable lengths, which has proven fine for almost all use cases in the past.