Re: [sqlite] updates, triggers and "out of order"

2004-07-22 Thread Daniel Franke
> No, because there usually is no inherent table sort order, unless you are > using a clustered index. Even then, I don't believe there is any > *guarantee* that the updates will happen in that order. > > You could, of course, use a cursor, retrieved in the order you want, then > update the recor

Re: [sqlite] updates, triggers and "out of order"

2004-07-21 Thread Brass Tilde
> > UPDATE work > > SET result = CASE > > WHEN in1 == 0 THEN 0 > > WHEN in2 == 0 THEN 0 > > WHEN in1 != in2 THEN 0 > > ELSE -1 > > END; > > I'll rephrase my question: > Is there a guarantee (maybe defined by SQL92-specs?) that in update statements > as the one above the rows are updated in table so

Re: [sqlite] updates, triggers and "out of order"

2004-07-21 Thread Kurt Welgehausen
> Is there a guarantee ... rows are updated in table sort order? No, theoretically a table is a set of rows; the rows have no inherent order. An implementation (like SQLite) may store rows in some order for convenience, but it's always free to change the storage scheme as long as the external beh

Re: [sqlite] updates, triggers and "out of order"

2004-07-21 Thread Daniel Franke
> > UPDATE work > SET result = CASE > WHEN in1 == 0 THEN 0 > WHEN in2 == 0 THEN 0 > WHEN in1 != in2 THEN 0 > ELSE -1 >     END; I'll rephrase my question: Is there a gua

[sqlite] updates, triggers and "out of order"

2004-07-20 Thread Daniel Franke
Hi sqlite-users :) I have found sqlite as the best solution to all my data storage problems - a great piece of software! Im currently working on a specific algorithm to fill in a table of numbers. Within the range of an index (idx), I have to count the numbers of rows where (in1==in2) and plac