Sort output

[thomas-reimonn]

This PR orders the results. By default, it's ordered by chrom, start, and id. Custom ordering can be provided using order_by.

If subsets of columns are selected, then the columns are sorted with available columns and finally by id.

[ivirshup]

I think we could do something like:

        order_cols = []
        if "gene" in tables:
            order_cols += ["seq_name", "gene_seq_start"]
        if "tx" in tables:
            order_cols += ["tx_seq_start"]
        if "exon" in tables:
            order_cols += ["exon_seq_start"]

to have a consistent total order. But idk for sure.

Can this get some tests for expected behavior?

[thomas-reimonn]

I updated it like that. It now sorts in a consistent order. I also added assertions to the tests to validate this functionality.

[ivirshup]

Could you add tests where we are selecting columns from a different table, e.g. .genes(["seq_name", "gene_...", "exon_id"]) and checking that the results are sorted?

[codecov-commenter]

Codecov Report

Attention: Patch coverage is 95.23810% with 1 lines in your changes are missing coverage. Please review.

Project coverage is 93.01%. Comparing base (ef33d87) to head (a5cdb15).
Report is 1 commits behind head on main.

Additional details and impacted files

@@            Coverage Diff             @@
##             main      #64      +/-   ##
==========================================
- Coverage   93.09%   93.01%   -0.08%     
==========================================
  Files           6        6              
  Lines         333      358      +25     
==========================================
+ Hits          310      333      +23     
- Misses         23       25       +2     

Files	Coverage Δ
src/genomic_features/ensembl/ensembldb.py	93.02% <95.23%> (-0.14%)	⬇️

[ivirshup]

Sorry for the delay on review/ approval but I wanted to dig into the duckdb behaviour we saw a bit more. An ibis maintainer believes this might be a duckdb bug.

• Order by columns not present in the result? ibis-project/ibis#8896