Rere: Another bug of IPLOAD !

Jake Salomon wrote: > > Valery, > > the purity of the afformentioned fertilizer is arguable. ;-) I know, I know, and I apologize. > While I would not have expected the result you got, hindsight tells me > it is reasonable. By default, ipload bypasses the whole ISAM layer, > writing pages directly to the tblspace. It skips index/btree > operations, skips constraint checking, skips any triggers.. you get the ^ You mean you can insert a NULL value in a NOT NULL column ??!!??? > idea. It is quite reasonable to assume it had skipped this step of > maintaining the serial value. ^ Now, this IS fertilizer... > > When the load is finished, the table is utterly unusable - you must > enable the database objects yourself and probably need to run UPDATE ^ yes, they are R/O until a level 0 archive > STATISTICS. (I taught the subject a few times but will still need a > refresher browse through the manual.) > > After you've done that, if the oncheck -pt still tells you the next > serial value is 1, *then* you'll have something to howl about. > AWWWOOOOOOOO!!! > I did the update statistics...I allways do the update statistics before rebuilding my indexes... Nevertheless, I did it *BEFORE* the level0...I'll check if it changes something. NOTA: I begun to have problems with ipload because I reorganized my 20 Giga instance last week... My examples are just...examples ;-) We matched the problem after 1 day of exploitation... I called the Tech Support, and they forwarded this "feature" in the US. > (In view of this, it occurs to me that your test would have been just as > valid if you had inserted 0-values 5 times and not bothered with the > skip in the serial sequence.) Yo ! I thought it was showing out the problem more obviously. > > -- Jake Salomon (Trying to stay out of the fertilizer business) Hi hi hi... > PS I am bcc'ing a friend at Informix to check my facts. K, can you > please post a response in case I have my info mixed up? Just about the FIRST "bug" of ipload: the Tech Support does not reproduce it on a Sun (they've got a -772 error...lucky men...). And the error occurs when allocating a buffer.... I'll give you more info when I have a complete answer from them. AWOOOOOOOOOOOOOOOOOOOOOOO !!! -- Valery Fouques valou@hotmail.com SYSADM GOSPORT-FRANCE /\\oO/\\ _ __/| \\`O.o' "355/113 -- Not the famous irrational =(_ _)= number PI, but an incredible simulation!" U PS: Sorry to ask, but WHO are you ?