In most places, and certainly in places that practice "common law" and democracy, part of what makes something legal is its moral acceptability. The laws on the books exist to reflect that and to make things function efficiently and fairly. Without that, laws are simply a tool of violence and oppression by the authorities (or the majority, in case of a democracy) against their population.
That's why we have juries of peers, and the ability of a jury to say "this person did the thing, but they shouldn't be punished for it".
Hence why jury nullification scares the ever loving crap out of the Justice system. Binding precedent is created by the Jury, which throws a wrench into the works that for some reason is seen less acceptable or "official" than a prosecutor exercising prosecutorial discretion.
That only makes things even more hilarious, because now the Justice system can't even claim to be consistently applying stare decisis across the board if that is the case.
Rather it only does it when someone makes the decision that to do so is convenient for maintaining the integrity of the Judicial system; thereby creating the facade that the entire thing isn't rife with capricious singularities like it actually is.
When laws are impossible to consistently enforce (as evidenced by prosecutorial discretion), or juries are not on board with seeing them enforced, it should be a much more blatant signal something is up or off than it is.
In fact, is there even a record of cases of "refused prosecutions"? If not, maybe there should be. Then there's be an objective metric to analyze to see if a law is being abused selectively.
Cases are recorded in general. I don't know if there's a BigQuery LexisNexis or whatever, but I bet someone has access to a representative database and can grep for "we the jury declare the defendant not guilty" followed by "the judge was foiled and slunk back to its lair to concoct a new scheme".
Not a ton to elaborate on, it just isn't precedent.
All jury nullification is is a jury finding someone not guilty despite the fact that they think they did actually commit all the elements of the offense. It doesn't prevent the law from being applied in future trials, even in identical situations (nor do other jury verdicts). It's not even typically known whether or not the jury found not guilty because of nullification or because they didn't find the prosecutions case convincing.
That's why we have juries of peers, and the ability of a jury to say "this person did the thing, but they shouldn't be punished for it".