I'm conducting a research project that consists in sampling different physical quantities on about 30 buildings, most of them single-family houses.

Some of those physical quantities will be common to all houses: indoor temperature, outdoor temperature, etc. Others, however, will be measured only on some houses. For example, we will measure the pulsed air temperature only on those buildings that are properly equipped with an air-handling unit. Such values make no sense for other buildings.

So we will have perhaps 50 different possible physical quantities. A given building will have, say, not more than 15 of these quantities.

One possible solution I'm considering is the Entity-Attribute-Value model. I would have one table for floating-point data (temperatures, heating water flow, etc) and another table for categorical data (water pump speed: high, low, off, etc).

The only concern I have is that the Wikipedia article mentions that this model is suitable for a "vast" number of possible attributes. This is hardly my case.

Any advice/recommendation/book/article would be most welcome.

  • 1
    50 possible factors of which no more than 15 are expected to be populated on a given entity qualifies as a "vast" number of possible attributes, IMO. This certainly is a sparse matrix problem. I would be very tempted to implement an EAV model. – Daniel Wilson Jul 11 '14 at 21:05

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.