how to get a pandas dataframe of ds.data_vars?

[raybellwaves]

I would like to turn ds.data_vars into a pandas.DataFrame with two columns: "Data variable" and "Dimension". This helps viewing datasets with a large number of variables. I can use pandas to export to csv, to filter for certain dims etc.

import xarray as xr
import pandas as pd

url = "https://thredds.ucar.edu/thredds/dodsC/grib/NCEP/GFS/Global_0p25deg/Best"
ds = xr.open_dataset(url)

ds.data_vars has a nice repr layout

ds.data_vars

If I plug this into a pd.Dataframe I just get the variables

pd.DataFrame(ds.data_vars)

I have some workable pandas code but didn't know if the DataVariables class had something built in or there are better ways to do this.

df = pd.DataFrame(ds.data_vars, columns=["Data variable"])
d = ds.variables
l = []
for _, row in df.iterrows():
    l.append(d[row["Data variable"]].dims)
df["Dimensions"] = l

[keewis]

I don't think we have anything like that, yet. For reference, it's much easier to compute all at once:

In [10]: ds = xr.tutorial.open_dataset("eraint_uvz")
    ...: pd.DataFrame(
    ...:     [[name, var.dims] for name, var in ds.data_vars.items()],
    ...:     columns=["Data Variable", "Dimensions"],
    ...: )
Out[10]: 
  Data Variable                           Dimensions
0             z  (month, level, latitude, longitude)
1             u  (month, level, latitude, longitude)
2             v  (month, level, latitude, longitude)