Quick Answer: Try a keyword search to find many citations and examples of the use of structural data alone to assist with column selection. Thousands have asked this before and you may be disappointed in the final results. There are many literature sources and databases where structure based method development may be helpful. Esp for unusual structures or special highly selective modes of chromatography (e.g. cyclodextrin columns). However, plenty of methods can not be refined based on structural data alone (it just is not that simple!) and logical, step-wise method development is needed to develop a good quality method. Often, we have no structural data for all of the compounds we are working with (or we are developing methods to identify a compound first). While column selection is one of the single most important parameters of method development, it is only one of many parameters which must be developed and optimized.
Column selection depends on the side groups and the retention mechanism. While a C18 column and 'reverse phase' would be my first choice, selection depends on the molecule.
I agree with Bruce, selection of the HPLC/UPLC column is dependent on the molecule and sample you are planning to separate/analyse. If the sample is hydrophilic in nature, column would be based on normal phase chromatography. On the otherhand, for non-polar sample, column would be reverse phase.