On Sat, 24 Jun 2000, George M Swan III wrote: > In other words a relational database optimizing performance by splitting > data into several tables to minimize storage space and other things. > Right??? Or is there something that keeps these two concepts separate: > 1) normalized and 2) relational ??? Actually, normalized databases are _slower_ than unnormalized database. Not sure about the storage space, but it's possiblely smaller (since you use an INT to refer to the ingredient instead of TEXT). A database is relational *because* it is normalized -- the "relation" refers to the references between the tables. > ... also I took note of the use of the term PRIMARY KEY, which would imply > indexing of the table upon which it was applied. Which in turn would make > the action of LOAD DATA INFILE take much longer according to De Bois and > that it's a balancing act to decide which would take less time, 1) > muscling through without removing the indexing or 2) removing or desabling > the indexing. And thus the importance of chewies' admonition, "[We should > first ask if the database is normalized or not.]" The use of primary key is basically necessary -- it allows you to refer to each element in the table uniquely. Thus you can say: SELECT * FROM recipies WHERE recipie_id="$foo"; or somesuch thing. Basically, you want indexes (via PRIMARY KEY, UNIQUE, and/or INDEX) on any column which you use in the WHERE part of a SELECT statement. It makes your queries faster at the expense of somewhat slower inserts. But for any table I've created, the selects are more important/more used than the inserts. Have fun. Luke --------------------------------------------------------------------- To unsubscribe, e-mail: tclug-list-unsubscribe at mn-linux.org For additional commands, e-mail: tclug-list-help at mn-linux.org