Joe discusses Hierarchical, Vector and Concatenation encoding before rounding up with general guidelines for designing encoding schemes.
Joe discusses how to deal with the kinds of encoding schemes, how to use them and how to design them. He discusses Enumeration, Measurement, Abbreviation and Algorithmic categories
Joe Celko introduces more powerful scales such as Interval, Log interval and ratio scales; before moving on to conversions, punctuation and units. Finally he gives guidelines as to how best to use scales in a database.
It is certainly possible to fake an Array in SQL, but there are only a few occasions when it would be the best design. Most often, the wish for an array in SQL is a sign of a forlorn struggle against poorly-normalised data. One of the worst sins against Codd is the repeating group, as Joe Celko explains.
Joe Celko discusses Nominal, Categorical, Absolute, Ordinal and Rank scales. These are the weakest scales we can use, starting with the weakest.