As this reaction is reversible, it is important to push the equilibrium to the right by removing the water generated. This can be achieved in one of two ways:
1. perform the reaction over the appropriately sized molecular sieves (i.e. so they only trap water, not diol)
2. Use Dean-Stark apparatus. The reaction should be performed in a solvent such as toluene at reflux, to drive off the water and trap it in the collecting arm.
The Dean-Stark method is the one most commonly employed. Use of a catalytic amount of acid will facilitate the reaction by both activating the carbonyl group to attack, and generating good leaving groups (i.e. water). A solid acid such as p-toluene-sulfonic acid is easy to weigh out, and can be removed by bicarbonate wash of the reaction mixture at the end of the reaction before concentration.
The aqueous wash should also remove excess diol (if you are using fairly small Mw diols), so it is fine to have a slight excess of diol present.
Failing this, depending in the polarity of your starting material, the idol can potentially be retained on the baseline in the correct eluent on a silica plug, so purification should be straightforward.
Monitor for disappearance of aldehyde (e.g. using 2,4-dinitrophenylhydrazine stain). Remember silica is acidic, so dry your silica/TLC plates well, otherwise it may look like your reaction is not complete when you run a TLC, if the plates have adsorbed atmospheric moisture.