I am just verifying that my assumption is right as I tried to see a qza tree in view.qiime2.org and got a 4 row table.
So: view.qiime2.org does not show phylogenetic trees because they are too large? And because there is not much to learn from them as the taxa IDs are nondescript?
I exported qza to newick and loaded it in
It is still interesting to see the clustering. Would be nice to have the sample IDs in the qza too, to see if OTUs cluster by sample, although an OTU can be linked to many samples.
Actually it has everything to do with the filetype. You are providing a .qza file (QIIME Zipped Artifact) which has no particular opinion on how to draw itself. Think of it as just a box of data and nothing more. This box still understands what kind of data is in it, and where it came from (provenance), so view.qiime2.org will still let you load these files so that you can look at those pieces.
If you need to visualize the data, you use a visualizer which will return a .qzv file (QIIME Zipped Visualization), these are things that have visual components to look at.
I like to think of it as .qza is data like .fasta meant for a computer; and .qzv is some human readable summary like .png or .pdf
You are describing a visualizer which would use both a Phyologeny[Rooted] artifact and a FeatureTable[Frequency] artifact. These artifacts are both useful in different contexts (and together) which is why they aren't a single entity.
That being said, we don't actually have a visualizer that does what you are describing in QIIME 2. If you are someone you knew was interested in making it, QIIME 2 is built to be extended with plugins, and we'd be happy to assist getting started in Developer Discussion!