Using spectroscopic databases for spectra prediction has become popular. Although most of the shifts predicted may be used as displayed there are some practical limitations. Some of these are suitable to distort results that are used without further consideration. Reasons and consequences of the most common aspects in this respect are discussed throughout this paper. Most critical effects are expected to rise from the presence of unrecognised data gaps and violation of the conventions used for structure coding for the database used. Any high standard deviation of predicted shift values or wide range of shift distributions indicate the need of further investigation. Looking up the reference data for the focus atom concerned can easily clarify most situations.