Does "version" in cat.access_nri still work?

Technical
1

In the release notes it says we can invoke previous versions of intake that uses the file_id option , as in
intake.cat.access_nri(version="v2025-07-30")
However, when I try that, I get the following error:

version=v2025-07-30 is less than v2026-01-16
Is this option still valid?

___
why am I trying to do this?
I have a student that is trying to use NCI intake to import OM2-IAF daily SIC.

she uses:

experiment1 = "01deg_jra55v140_iaf_cycle4"

access_var_ice = 'aice'

access_variables_m = catalog[experiment1].search(variable=[access_var_ice],frequency="1day").to_dataset_dict()

and then she extract the keys she wants…

But this takes a very long time for just 2d variables (Almost 10 mins), because .to_dataset_dict() is accessing two keys and later on extracting the one we want. the two keys are:
['seaIce.1day.d2:2.nc:5.ni:3600.nj:2700.mean', 'seaIce.1day.d2:2.nc:5.ni:3600.nj:2700.nkice:4.mean']
so I wanted to reduce that by making the code access right away just the file_ID we want

2

@Wilton_Aguiar @dougiesquire should have a look at it this afternoon, hopefully.

4

Hi @Wilton_Aguiar. We might have to wait until @CharlesTurner is back from leave to answer your question about using old versions. But I can help with accessing the data you want using the latest version.

The reason this takes so long is because these data include the 2D grid coordinates TLON, TLAT, ULON, ULAT in every single file. With its default arguments, xarray works hard to check that these coordinates are compatible at every time. This has come up a few times before, e.g. in the context of the cosima cookbook here.

You can pass some additional arguments when you open the data to tell xarray not to do these checks. This will speed things up:

experiment1 = "01deg_jra55v140_iaf_cycle4"

access_var_ice = 'aice'

esmds = cat[experiment1].search(variable=[access_var_ice], frequency="1day")

xarray_combine_by_coords_kwargs = dict(
    compat="override",
    data_vars="minimal",
    coords="minimal",
    join="outer",
)

access_variables_m = esmds.to_dataset_dict(
    xarray_combine_by_coords_kwargs=xarray_combine_by_coords_kwargs
)

Or, if you want to filter based on the file_id before opening the data you can do something like:

file_id = 'seaIce.1day.d2:2.nc:5.ni:3600.nj:2700'

access_variables_m = esmds.search(file_id=file_id).to_dask(
    xarray_combine_by_coords_kwargs=xarray_combine_by_coords_kwargs
)
5

thanks @dougiesquire , that reduced importing from 10 mins to 7 secs! :rocket:

6

@Wilton_Aguiar if you’re happy with my response as a solution to this topic could you please mark it as “the solution” by clicking the little tick at the bottom of the post?

8

I think the version stuff is related to our updates that wrap the catalog in different object to let people alias common CMIP names. I’ll take a look into it.

Just a heads up that using previous versions of the catalog has somewhat limited utility, as they still have to refer to the same underlying files on disk. Typically this means using an older catalog is more likely to give you stale file handles which were updated in a newer version.

Also, worth being aware (if you weren’t) that you can also explore this datastore via
interactive-data-catalogue/01deg_jra55v140_iaf_cycle4 which should be easier for a new user.