Before starting a new research project/paper, one needs to make sure that this project is new and nobody has done it. I'm still not confident about whether sth really hasn't been done in the literature by doing the following search: 1.google scholar, 2.researchgate, 3.related papers, their reference lists and what papers cited them. Is there a foolproof approach?
Ideally, your advisor would know well beforehand if something is novel or not. In practice, this is not always the case.
The best thing to do if you want to develop new stuff if to build upon new stuff as well. If anyone could have done what you are doing 20 years ago, then you'd need to check 20 years of literature to be sure, but if you are extending the results of a known paper which was innovative in itself and published on a respected journal 3 years ago, then you can basically just check the references quoting it.
Also, try to have some preliminary work published on an international conference rather than a journal. This will give you the opportunity to present your work to an audience of both students and established researchers in your field, see their works and to talk to these people over coffee breaks and social events. Someone is likely to point to you if you are missing something important and known in the literature. They'll likely give you new terms and names which you should look for, I've seen a few topics which don't really have consistent nomenclature throughout publications.
Then again, if some result which is relevant to you actually appeared in some obscure publication (which maybe is not even in english), that nobody knows or quoted, then for the purpose of your question, why does it matter?