there really isn't such a thing as a panel memory style, a panel memory points to a pattern, either preset, custom or composer, and points to a sound(s) either preset, EW board, or sound memory, it doesn't contain them. To create a style you either keep the sound(s) and point the panel memory to a new pattern, or vice versa, whichever is easier.

Thus if your disk A is a preset pattern, just load the panel memories from disk B, and resave each panel memory pointing to that same preset pattern. If the sounds depend on user edits load the panel memory and sound memory. If disk A is a pattern in a composer, load file B and then load the composer only from Disk A, then resave each panel memory pointing to the correct composer and variation etc. Use page 2 of the load page to select which subfile (composer, panel, etc) to load.

If you have individual composer variations or sound edits to superimpose into existing composers and sound memories use page 3, single load, of the load page.

The song answer depends on whether the song is an easy record, a realtime multitrack or a midi file, and whether the sound is a preset or an edit. If an easy record with a preset sound it could be as simple as resaving a panel memory and perhaps doing a panel write. If a multitrack or midi it could mean substituting the sound in step record. If a sound edit, it could mean loading the sound memory first. If other sound edits are used in the song it could mean single loading the individual edit to an unused slot so as not to overwrite the other sound edits used in the song. Thus many answers depending on the exact song