--two-pass flag only uses two sequnces in sort

[Snerken]

Prerequisites

I am using v0.9.1-dev and have read "seqkit sort --help"

Describe your issue

I am following the examples given in the chapter on SEQKIT in the Biostar Handbook Bioinformatics course.
When I use the --two-pass (-2) flag with the sort function the program just returns two sequences. This does not happen when the two-pass flag is omitted.

Example with two-pass flag:
Fri Oct 05 snerx@kali-linux:~/BioPython/biostar-handbook-September-2017/Lectures/seqkit_fastq
$ seqkit sort --by-length --two-pass viral.1.1.genomic.fna.gz > viral.genomic.sorted_2pass.fa
[INFO] read and write sequences to tempory file: viral.1.1.genomic.fna.gz.fastx ...
[INFO] create and read FASTA index ...
[INFO] create FASTA index for viral.1.1.genomic.fna.gz.fastx
[INFO] read sequence IDs and lengths from FASTA index ...
[INFO] 2 sequences loaded
[INFO] sorting ...
[INFO] output ...

Example without two-pass flag:
$ seqkit sort --by-length viral.1.1.genomic.fna.gz > viral.genomic.sorted.fa
[INFO] read sequences ...
[INFO] 7 sequences loaded
[INFO] sorting ...
[INFO] output ...

Seems like a bug of some kind to me.

Regards

TEF
Bodø
Norway

[shenwei356]

Hi @Snerken , it is a BUG existed for almost 2 years. I'm so sorry.

Affected commands include sort, shuffle, split, it occurs when using -2/--two-pass to process .gz files.

Fixed now: v0.9.1-dev2