aglew@mcdurb.Urbana.Gould.COM (02/26/89)
Maybe someone in comp.databases can help me out. I am not very experienced with databases. I wish to have a "database" (in the general sense - not necessarily an RDBMS or the like) for various performance measurements I take on various computer systems. Each data point is conceptually a point in a the space of possible measurements X configurations; eg. MEASUREMENT_VALUE = 140 MEASUREMENT_UNITS = kB/s MEASUREMENT_PROGRAM = "disk performance test" MEASUREMENT_PROGRAM_VERSION = ... with configurations SYSTEM = System V/68 v. XXX HARDWARE_CHASSIS = 1147 etc., etc. My problem is that I *do* *not* know in advance what are the important fields or classifications; eg. a new board may come out with a completely new configuration option that I need to record; or, I may discover that a factor I was unaware of needs to be considered in all future measurements. With my home-brew record system, I handle this by making every measurement a keyword oriented tuple, with provision for handling keyword/tuple elements that are not present when I process the records. I'd like to use a commercial database system, but every time I try the first step seems to be "define the fields of your records". And I can't. Even if I did, my records will be highly irregular, and any set of fields I define will have a lot of fields marked as non-applicable to any particular measurement. Usually I end up with a generic "coments" field in the record - with most data in the comments. If I do that, what is the purpose of using a database? Are there databases that are structured the way I want? Or, are there ways of using conventional "fixed record format" databases the way I want? Please send email. Andy "Krazy" Glew aglew@urbana.mcd.mot.com uunet!uiucdcs!mcdurb!aglew Motorola Microcomputer Division, Champaign-Urbana Design Center 1101 E. University, Urbana, Illinois 61801, USA. My opinions are my own, and are not the opinions of my employer, or any other organisation. I indicate my company only so that the reader may account for any possible bias I may have towards our products.