if action is invariant under a translation x->x+a then there is a conserved current. I was wondering how should we know that if action is invariant for a particular translation in the first place. I know the only way which is when Lagrangian has some canonical variable absent, but this does not look very appealing to me for all the situations. Is there a more encompassing understanding. kindly share your understanding.