My March post @All Analytics.
Frequent hits @dbdebunk.com are driven by the question “Are keys
mandatory?” Puzzlingly, many data professionals do not seem to
understand why duplicates should be prohibited. This should worry
analysts. But “Stating the same fact more than once, does not make it truer, only redundant,” as E. F. Codd used to say. The absence of an identifier means that individual entities are
not meaningful, so this representation contradicts the real world. Contradictions produce problems. First, a DBMS is incapable of
“visually” discerning a data entry duplication error from "valid"
duplicates, which means high risk of inconsistent databases and wrong
counts and other query results.
Read it all
Please comment there, not here.