All: Confession.
I jumped off V in mid-90s after several years of gainfully brute-forcing masses of raw data from the R> prompt. Never found the time to actually program anything or develop front end. Since then have produced schema (in RB4K notation) for relaxation and sanity, but not one stick of code. Now, a deep reservoir of of disjoint, un-Coddsian data threaten to breach the FileMaker/Excel dam and swirl me and mine into non-normalized oblivion. My arm-chair schema may not survive the deluge. But a man's gotta try. Interviewed a number of you (thanks all!), bought v8; been blown away by the incredibly giving R-List culture and the deep, deep integrity of the Razzak team, the product and the community you have built. But, holy socks! So much to learn. My plan is to work up to the point where - like, if ever - I can contribute as cogently as do you all. In the meantime, try not to waste anyone's time, and be ever thankful for my daily RBase-L fix. And risk coming off as the smart-aleck-in-residence. To business. A schema question regarding 'menage-a-many-many'. Note: '[' ']' '[pk]''[fk]' ':' used for disambiguification only. ****** MODEL ****** Logical Relation #1: part <|---o>> part2assy <<o---|> assy Physical Relation #1: part:[pk]partid <|---o>> part2assy:[fk]partid assy:[pk]assyid <|---o>> part2assy:[fk]assyid ...resulting in Compound Primary Key: part2assy:[pk][partid + assyid] Ok, that's fine. But table 'part2assy' has a many-many relationship with table 'fabprocess' through join table 'fabevent': Logical Relation #2: part2assy <|---o>> fabevent <<o---|> fabprocess Physical Relation #2: ... ****** Question: ****** a) Can a compound PK be at the 'one' end of a relationship? ex: part2assy:[pk][partid + assyid] <|---o>> fabevent:[fk]umgeeIdontknow or: b) Should I create an auto-numbered stand-alone PK for part2assy ex: p2aid part2assy:[pk]p2aid <|----o>> fabevent:[fk]p2aid ...and take up the slack with 'AND' entries in the WHERE clause? Thanks all bruce/safesectors
