Sony does shenanigans like that.
Canon does not.
Whether you want to stick to a pure ideal of 'Raw', or accept some flexibility when it is useful, is of course up to you.
Please note that all 'Raw' formats are not directly the 'true' sensor output, actually the sensor output is processed to produce a standardised and usable image format, in most cases compressed, either lossless or lossy to keep file size down. Then the question arises what level and kind of processing might be allowed for a 'Raw', but in any case it is already processed according to the OEM's intended output, while the processing details are not disclosed and essentially depend on the OEM's decision.
The usefulness of 'Raw' videos/images is that important parameters like white balance, gamma, noise reduction etc, can be selected in post and lossy compression is normally avoided or minimal to make a maximum of image information available for post-processing.
In the case of Sony's Burano camera oversampling 8K Raw to 6K 'shenanigans' Raw, all the useful items of a 'Raw' file are preserved except for the intended reduction in pixel dimensions and file size, while there is a significant gain in image quality thanks to the oversampling, in particular due to averaging out noise and increasing (colour) resolution when the sensor uses the typical Bayer pattern (like almost all commercial cameras).
The benefit of oversampling is very obvious when comparing EOS R5's standard 4K to 4K HQ, which is oversampled from 8K.
A drawback of oversampling is the required processing power in camera causing battery consumption and heat generation, while current top camera CPUs, including the EOS R5 for 4K HQ (not available as Raw), already achieve it, and future generations will get more and more efficient.
In conclusion, to me oversampling of video makes good sense for Raw and any other video format.
When oversampling of video is applied, then e.g. 60MP produce significantly better quality than 45MP for video and for stills.