2 ms·
And how much overhead in logic, code and frustration would that cause in terms of development and support.. right now, I'm dealing with an over-normalized datab
by tracker1 14y ago
And how much overhead in logic, code and frustration would that cause in terms of development and support.. right now, I'm dealing with an over-normalized database close to what you are describing and needing over 20 joins in a single query to get a complete record for display (not including actual sub-records) but to get a complete set of properties, where null means not there...
- dragonwriter 14y ago> And how much overhead in logic, code and frustration would that cause in terms of development and support.. Depends on the competencies of the people doing dev and support. Personally -- both as a developer and a technical user -- I've had more problems dealing with situations where NULLS had ambiguous semantics, where the typical naive use of nullable columns instead of normalization into logical units of data that must all be present or absent together resulted in avoidable data inconsistencies, etc., than I've ever had with overnormalized tables. Joins for queries are a solve-once development problem; data inconsistencies and ambiguities resulting from the problems with NULL are an ongoing problem.