Strange behavior in feature-table summarize

[Arron_Shiffer]

When using qiime2-2018.2 and performing:
qiime feature-table summarize
--i-table EXMP-200-single-table.qza
--o-visualization EXMP-200-single-table.qzv
--m-sample-metadata-file ../EXMP_Sample_metadata_3_2018.tsv

The feature table qzv has some strange sorting behavior see the below pics:

It first sorts by sample ID while they are all fully numeric. It then begins to sort by sequence count when the sample ids become alphanumeric.

[Mehrbod_Estaki]

Hi @Arron_Shiffer,

By chance was this feature table the product of two feature tables merged? I noticed a few versions ago that when I merged 2 tables the sorting didn’t follow the usual Highest reads >> Lowest Reads format anymore. I totally forgot about reporting that till I saw your post.

[Arron_Shiffer]

Hello Mehrbod,
These come from one feature table (not merged).
Thanks,
Arron

[thermokarst]

Thanks for the reports, @Arron_Shiffer & @Mehrbod_Estaki.

As you mentioned @Arron_Shiffer, it looks like the table is being sorted by Sample ID, rather than Sequence Count. This looks like it is a result of how pandas handles numeric identifiers. I have opened an issue on the relevant issue tracker to put a forced-sort on our radar. We will update this issue when an enhancement is released. Stay tuned!