Fix the SQL dialect bug in every dialect

01 · Before and afterThe pattern

One tokenizer for every dialect instead of string concatenation per driver.

// before (in several dialect files): const jsonPath = '$' + userPath;
export function toJsonPath(userPath: string): string {
  return tokenize(userPath)                     // e.g. lodash toPath
    .filter((part) => part !== '')              // ".a..b." → ["a","b"]
    .reduce((acc, part) => /^\d+$/.test(part) ? `${acc}[${part}]` : `${acc}.${part}`, '$');
}

02 · Run it on your appThe check

If your app supports more than one database, does every dialect run the same input-normalisation code, with tests that run on each one?

If the honest answer is no or not sure, this is worth a look before it reaches production.