I am using v2.0.6 of featureCounts on a Linux machine. I am running the following command:

$ subread-2.0.6-Linux-x86_64/bin/featureCounts -a fc_test.saf -F SAF -o fc_test.counts.out -T 2 -p fc_test.bam

This returns an output where none of the alignments are assigned. That's because I have the following in my SAF vs. BAM:

## SAF
$ head -n2 fc_test.saf
gene-NALCN      NC_086317.1       334712972       334713283       +
gene-NALCN      NC_086317.1       334786484       334786618       +

## BAM
$ samtools view fc_test.bam | head -n 1
SRR5119615.4927416.1    99      NC_086317       334624765       0       73M642546N17M   =       335446699       822024  CCTAAATCCCCTTCCTCAGATGTTGAGGTTCGGACTCTATCAAATATTCTGTACTCTGTCCCTTTTTTTTTTTACTCCCAAATTCAATTT*       NH:i:2  HI:i:1  AS:i:143        nM:i:15 NM:i:11 MD:Z:10T14G4A32G0G0G5A0C1G0G0T13        jM:B:c,1        jI:B:i,334624838,335267383      XS:A:+  MC:Z:90M

My BAM and SAF files use different seq-ids for the chromosome. So, I included the following alias file:

$ cat aliases.txt 
NC_086317.1     NC_086317

And run the following command:

$ subread-2.0.6-Linux-x86_64/bin/featureCounts -a fc_test.saf -F SAF -o fc_test.counts.out -T 2 -p -A aliases.txt fc_test.bam

Still, I see that none of the alignments have been assigned. On the other hand, if I modify my SAF file to remove the version number from the chromosome seq-id, I see ~25% of the alignments assigned.

I think this is a bug in featureCounts and present at least in an older version v2.0.2 and perhaps even earlier. I read in an older post that this could be because of line endings. In my case, all line endings are unix-style.