Using ODB data

import earthkit.data as ekd

ekd.download_example_file("test.odb")

ODB is a bespoke format developed at ECMWF to store observations. The following code reads some ODB data with from_source() from a file containing forecast and analysis departures:

d = ekd.from_source("file", "test.odb")
d

ODB file

---

path	test.odb
size	13 KiB
types	pandas

to_pandas() converts data into a Pandas dataframe (uses pyodc under the hood):

df = d.to_pandas()
df

	lat	lon	fg_dep	an_dep
0	38.808998	4.292600	0.514543	0.513129
1	73.693100	-3.416700	-0.098977	-0.123831
2	31.168501	-29.865499	0.420131	0.389983
3	49.040798	-19.877501	-0.182703	-0.124131
4	73.351097	40.176102	0.087687	0.149949
...	...	...	...	...
712	45.798801	-0.527100	0.103859	0.132765
713	73.788597	-8.806200	-0.069707	-0.080577
714	60.162998	-14.572900	0.138816	0.011107
715	53.455299	-29.641600	0.082582	0.184872
716	40.669201	29.702101	-0.023861	-0.032721

717 rows × 4 columns

We can directly pass arguments to pyodc. E.g. columns specify a subset of the ODB columns we want to extract:

df1 = d.to_pandas(odc_read_odb_kwargs={"columns": ("lat", "lon", "an_dep")})
df1

	lat	lon	an_dep
0	38.808998	4.292600	0.513129
1	73.693100	-3.416700	-0.123831
2	31.168501	-29.865499	0.389983
3	49.040798	-19.877501	-0.124131
4	73.351097	40.176102	0.149949
...	...	...	...
712	45.798801	-0.527100	0.132765
713	73.788597	-8.806200	-0.080577
714	60.162998	-14.572900	0.011107
715	53.455299	-29.641600	0.184872
716	40.669201	29.702101	-0.032721

717 rows × 3 columns