On trying to find $\displaystyle{\frac{dy}{dx}}$ when the given equation is $\displaystyle{xy = \cot{xy}}$, the following step gives an equation which gets rid of the derivative as: $\displaystyle{\frac{d}{dx}(xy) = - \csc^{2}{(xy)} \frac{d}{dx}(xy) }$
what's more, it gives an equation with no $(x,y)$ in real plane satisfying the equation, as: $\displaystyle{\csc^{2}{(xy)} = -1}$
To be sure that the function isn't actually differentiable, I graphed it on desmos, it looks although discrete at some points away from origin but I take it as inability of the graphing software to plot precisely, also supported by the seeming smoothness of the curve near origin.
So should I consider it non differentiable and leave the problem there or is there just any mistake i did while differentiating, and the curve is actually differentiable?


The equation $$t=\cot t$$ has infinitely many solutions, let $t_k$.
So
$$xy=\cot xy$$ describes infinitely many equilateral hyperbolas
$$xy=t_k.$$
Every branch is differentiable ($\dfrac{dy}{dx}=-\dfrac{t_k}{x^2}$), but collectively, $y$ is not a univocal function of $x$.
If you want to use implicit differentiation,
$$xy=\cot xy$$ yields
$$y+xy'=(-\csc^2xy)\,(y+xy')$$ which is
$$y'=-\frac yx=-\frac{xy}{x^2}.$$