Re: SE7.23: using today +- n question

Jonathan Leffler wrote: > > Richard harnden wrote: > > > I had this problem too. > > > > Tried: INSERT ... VALUES(DATE(today - 3)) > > INSERT ... DATE(EXTEND(CURRENT, ... > > > > Nothing Works. (which is normal for my sql, btw :) > > > > Ended up having to insert TODAY, then: > > UPDATE ... SET (datum) = (datum - 3) WHERE datum = TODAY > > > > Annoying. > > > > >>> Peter <proORGA-Leimen@t-online.de> 03/03/00 02:54pm >>> > > using "select today - 3 from sysusers", works as expected, > > using "insert into A (datum) values ( today) works again, > > why will not work "insert into A (datum) values ( today - 3) ? > > Ah, the hysterical raisins strike again! > > Once upon a long time ago, there was a little computer manufacturer > called IBM. They had a few people to spare, so they sent off Ted Codd > and he invented relational databases. They still had a few people to > spare, so they sent them off to create System/R and SQL, to implement > an approximation to Ted's relational databases. > > They also invented a language called Embedded SQL for use within > programming languages -- their language of choice was PL/1, with Cobol > lurking in the wings too. None of the single-letter nonsense like C, > of course, and no caffeinated languages either; they were far in the > future. > > Anyway, to help programmers, they invented a syntax for inserting a > set of values from a set of program variables into one row of a table: > > INSERT INTO TheTable VALUES(:var1, :var2, :var3, ..., :varN)> > Now, it would be fiddly to work out how to evaluate expressions > involving those values down in the DBMS, and the language could do that > quite happily and even quicker, so...so we'll insist that only simple > values are allowed in the VALUES list. No expressions. > > And lo, System/R was deemed a success, and was transformed into DB2, > and became an ANSI and ISO standard, and no-one saw fit to fix the > non-orthogonal nature of the values listed in the VALUES clause of > an INSERT statement, even though luminaries such a C J Date (on whom > be peace) pointed out the detrimental effects of such mis-features. > The most noticeable such effect being the confusion it causes people > new to the language who reason thusly: > > It works in this context; it works in this different context; this > works and looks very similar -- why the blazes doesn't it work? > > Said people are then dreadfully disappointed when a cranky geek > points out that unfortunately, their plausible logic is flawed > because of hysterical raisins -- sometimes misspelled as historical > reasons. > > :-( I bet your little'un enjoys bedtime stories with YOU! :-) ... and Snow White said to the seven little computer hackers, can anyone hack into the Queen's MirrorMate for me? I'd like to alter it's expression handling routines ... :-O Cheers, -- Mark. +----------------------------------------------------------+-----------+ | Mark D. Stock mailto:mdstock@mydas.freeserve.co.uk |//////// /| | http://www.informix.com http://www.informixhandbook.com |///// / //| | http://www.iiug.org +-----------------------------------+//// / ///| | |What year 2000 bug? year 2000 bug? |/// / ////| | |year 2000 bug? year 2000 bug? year |// / /////| | |2000 bug? year 2000 bug? year 1900 |/ ////////| +----------------------+-----------------------------------+-----------+