Filter duplicate taxIDs by order/class?

[philippbayer]

Prerequisites

• [X] make sure you're are using the latest version by taxonkit version
• [X] read the usage

Describe your issue

As described in the manual for name2taxid, some names like 'Drosophila' return several taxonomy IDs. Would it be possible to add a flag (either in name2taxid or filter) that retains taxonomy IDs if they are in a specific class or order? In my case I'm only interested in fishes, so if the other duplicates are from plants etc. I could just remove them.

For example,

echo Rondeletia bicolor | taxonkit name2taxid | taxonkit filter --tax_subset 7898

would keep only 1311492 and discard 1368176 (as 1311492 is 'within' Actinopterygii, 7898). I have yet to encounter duplicate taxonomy IDs within the same class/superclass.

[shenwei356]

Currently, you can filter the result using a white list from taxonkit list.

$ echo Rondeletia bicolor \
    | taxonkit name2taxid \
    | csvtk grep -Ht -f 2 -P <(taxonkit list --ids 7898 -I "")
Rondeletia bicolor      1311492

[philippbayer]

Thank you!!! That is good enough for me; I only work wit hfish so it's usually 7898, sometimes 7777. I'll close the issue

[philippbayer]

After implementing this filter I found 6 fish that have duplicate taxonomy IDs where both IDs are within the fish.

It seems that for those, taxonkit returns the 'good' taxonomy ID first followed by the outdated one. So I did this to get rid of the second hit using awk to keep only the first taxonkit-return per query ID:

grep '>' BOLD_chordata_28_07_2023.namesOnly.COI_species_only.fasta | sed 's/|/\t/g' | taxonkit name2taxid -i 2 | awk -F'\t' 'NR==1 || !a[$1]++' > Taxids.txt

These are the species in my dataset, I'm sure there are others: Bryconamericus caucanus, Centropogon australis, Oreonectes furcocaudalis, Pelvicachromis signatus, Trichomycterus variegatus.
But it works so I'm keeping this closed for others in the future :)