@James Fellows Yates The pipeline finished successfully when building a Diamond and Kraken2 database with the Viral Refseq sequences (downloaded from https://ftp.ncbi.nlm.nih.gov/refseq/release/viral). Here is my samplesheet and command line (nextflow v24.10.5)

id,taxid,fasta_dna,fasta_aa
ViralRefSeq,1,viralrefseq/viral.1.1.genomic.fna,viralrefseq/viral.1.protein.faa

nextflow run nf-core/createtaxdb -r 1.0.0 \
  -c htc_resourceLimits.config \
  --input samplesheet.csv \
  --accession2taxid /bgfs/projects/nextflow/nfcore_createtaxdb/ncbi_taxonomy/nucl_gb.accession2taxid \
  --prot2taxid /bgfs/projects/nextflow/nfcore_createtaxdb/ncbi_taxonomy/prot.accession2taxid.gz \
  --nodesdmp /bgfs/projects/nextflow/nfcore_createtaxdb/ncbi_taxonomy/nodes.dmp \
  --namesdmp /bgfs/projects/nextflow/nfcore_createtaxdb/ncbi_taxonomy/names.dmp \
  --build_diamond \
  --diamond_build_options='--no-parse-seqids' \
  --build_kraken2 \
  --dbname vrs \
  --outdir results \
  -resume

Then I ran diamond using a viral test sequence and it successfully returns the staxids for the hits:

diamond blastx --query ~/src/data/tag.cdna.fa --db results/diamond/vrs-diamond.dmnd -f 6 staxids stitle sscinames pident evalue

1891767	NP_043127.1 large T antigen [Betapolyomavirus macacae]	Betapolyomavirus macacae	100	0.0
1236395	YP_009111344.1 large T antigen [Betapolyomavirus cercopitheci]	Betapolyomavirus cercopitheci	75.2	0.0

Tested Kraken2 and it works

kraken2 --db ../results/kraken2/vrs-kraken2 --output kraken_output.txt --report kraken_report.txt --use-names ~/src/data/tag.cdna.fa

cat kraken_output.txt 
C       tag.cdna        Betapolyomavirus macacae (taxid 1891767)        2127    1891767:70 0:38 1891767:17 151341:5 1891714:2 1891767:80 0:29 1891767:134 1891714:5 1891767:64 151341:5 1891767:75 1891714:5 1891767:431 0:42 1891767:217 0:38 1891767:70 1891714:2 1891767:151 151341:1 1891714:5 1891767:52 1891714:7 1891767:548

So, it seems that any generic TAXID (

1

in my example) that is put in the samplesheet works at least for Kraken2 and Diamond when using one file containing ~18,000 sequences. Thank you for your help.