Spectral surveys in the past were a hobby of a few, usually restricted to strong, line-rich and close-by sources which were considered templates for source classes, e.g. Orion KL for hot cores, IRC+10216 for AGB stars, and CRL618 for protoplanetary nebulae. Not any more, since with the large bandwidths and high sensitivities of modern instruments, notably ALMA, all but a few sources will show many lines from many molecules at every observations. So (involuntary) line surveys will be the norm rather than the exception. A common strategy is to ignore all lines but the few one is interested in. Since all data will be available through the archive, this does not mean that the data are lost, since eventually the information will be extracted. Another strategy is to take the bull by the horns, and try to analyze all or at least a large portion of the spectrum. This includes the steps of line identification, source modeling and linking to physical and chemical models. With the data volumes at hand doing it the traditional, pedestrian, way is somewhere between impractical and impossible, semi-automatic methods need to be employed.