How to match the coord's of a PSD layer in Resolve 15?

I very recently started using DaVinci Resolve 15, and am growing to love it, but I've now hit a bit of a rookie wall.

I have used Affinity Designer, to create a layered video template. Now, since there is no interoperability between Serif and Blackmagic's software, I'm forced to rasterise all layers separately and import them one by one into Resolve. And this is where my problems lay.

Back when I used Photoshop and Premiere/After Effects, I could simply import a PSD and choose the layer size mode. When I picked "retain layer sizes", the layers lost their positioning, but retained their sizes in proportion to one another. From there, I could just take each one, and copy their X/Y coordinates from Ps to Pr/Ae and they would return to their exact places.

How do I replicate that in Resolve?

I'm importing the rasterised layers successfully, but when I use the Inspector, it appears that Affinity Designer and Resolve are using two different coordinate systems, so I can't return layers to their original locations manually.

If I centre a layer in Designer, the transform panel shows me the exact X/Y coordinates... In this case, for a 1920x1080 file, the centre is, unsurprisingly, at X-960px/Y-540px... But centring in Resolve just outputs X-0/Y-0. How do I work around that?

Well, it's simple, as long as the format is the same:

Affinity is using the Adobe model of counting from the upper left corner, and Resolve is setting 0 at the center.
If you change the timeline format, pixels values become related to that format, which seem to make things complicated if you later change the format.

IMHo, the best solution is in Fusion, where everything is a percentage, so it's resolution independent.

I've since transitioned to Fusion for this project, but the result is the same.

I was given advice on reddit to transform pixel based coord's to float-based ones by dividing the X pixel value by my project's height and Y pixel value by the project width. This should, by all means, work. However, for whatever unfathomable reason, and I sincerely hope this is a bug and not an intended feature, the transform node moves the object in relation to said object's own size. Not to the size of the composition.

So whenever I move said object to X-0.0, it just goes to where its left-most edge used to be when it was centred.

What I need it to do is, once set to X-0.0, to put its anchor/swivel point, which is at the centre of the object, on the left edge of the composition. This way I'll be able to translate the coord's correctly, as they won't be relative to the object being transformed, but rather to the size of the whole composition.

I'm not brilliant at explaining, so here's the visual aid.

This is happening in Resolve 15, in the Resolve-integrated Fusion, AND in the standalone Fusion 9.

Well... I found a workaround by using the merge node's X/Y adjustment inputs, which work as intended, but the problem with the transform node still persists.