Domains, Lists And Data
Mastering data is all about mastering the basics, and there is no better place to start than domains. A domain is basically a list of like-minded items. It can be as simple as colors, or as complex as chemical compounds. What makes it a domain is that all the items fit within the definition of the domain. A domain can have as few as one entry to as many as billions of entries.
Some domains are just a static set of rows, others may auto-extend as you add new values. There are some special cases; numbers and dates. For numbers we just pick a valid subset - as numbers can range from -infinity to +infinity. It is the same with dates as well (they are just a different form of numbers).
There are a lot of standard domains - that are common across most existing databases. In theory, it should be easy to obtain any standard domain for a new project you are working on. But, in reality, many organizations will try and monetize the standard domains - the ease of getting a verified domain can be attractive to a lot of people. But, if you keep looking, the world has many kind souls.
A lot of domains will just live within your corporate world; the business terms, acronyms, transaction type, transaction status. You set the values and define the standards.
Building records is just about selecting the attributes for the table. Each attribute has an associated domain. You then select the values for your record and write them to the record. In the pure world of data you actually don't persist the values you just record the intersections of the domains. But, to ease usage of the values, every modern database records the values and not the domain positions.
领英推荐
If you think about it simply, then a record is just taking entries from multiple lists. That is relatively easy to imagine, you just slide the lists up and down until the values align.
Visualizing domains is a personal choice. For a simple system it can look like the wheels from a one arm bandit. For a complete eco-system it might look like star scape with domains intersecting and overlapping. It comes down to what you are comfortable with.
There are always opportunities to push the envelope for data. AI is not about making what we do better, it is about reinventing the storage and usage of data.