Greetings qiime2 developers,
I’ve got a dual-barcode dataset that includes multiple genes (16S, 18S, ITS) together on a single MiSeq run (it is becoming more common these days). I’ve managed to demultiplex these samples in qiime2 after some days of head-scratching. Now, in order to process some of these downstream (stuck at the phylogenetic methods because it only takes rep-seqs.qza), I need to split them (since ITS can’t be aligned). This thread Method to filter demultiplexed file was very helpful. However, I was hoping that I didn’t have to go through and demultiplex each gene separately, adding to the computation time on larger datasets.
So, my suggestion here would be that you good folks create a way to split the demux.qza file (perhaps by the various methods highlighted in your feature-table filter tutorial) in a future qiime2 release. That would add more flexibility to the whole system. 
Many thanks!
Nhu