Read NetCDF files from S3 bucket without downloading the files in the local system

[najiya-12]

I am looking for one year of ABI-L2-CODC data, however, it will be a very large volume of data using goes2go.
data for one month is around 8000 files

[blaylockbk]

Hi @najiya-12, thanks for your question.

goes2go doesn't currently have this capability and I'm not currenlty planning to implement this (pull request is welcome!).
goes2go can give you a list of file paths for a time range:

# note: this uses the GOES class from the most recent push to the main branch
from goes2go import GOES
G = GOES(16, "ABI-L2-COD", domain='C')
df = G.timerange('2021-01-01', '2021-02-01', download=False)

With that info, xarray does have some capability of reading data directly from S3 without downloading it (see here). (I'm run into errors when I do that ValueError: I/O operation on closed file.; I wonder if that's an issue with my virtual environment or if it's a window problem.)

---

What you are trying to do sounds like a "big data" problem. I would recommend checking out some other tools like https://github.com/fsspec/kerchunk. They might have the right tooling for your needs.