Re: AW: [sqlite] primary key and physical data organization

2006-04-05 Thread Dennis Cote
Martin Pfeifle wrote: Thanks, If we look at a B+-tree, all records are stored at the leaf level and form at least a logical list. But what happens, if we insert new records which do not fit on the corresponding leaf page any more. Assume this page has to be split. Where is the newly created p

Re: AW: [sqlite] primary key and physical data organization

2006-04-05 Thread drh
Martin Pfeifle <[EMAIL PROTECTED]> wrote: > Our goal is that all records are physically clustered according to their RO= > WID. > In order to achieve this goal, does it make sense to reorganize a table by = > for instance a command like > "insert into reorganized_table > select * from original_tabl

AW: [sqlite] primary key and physical data organization

2006-04-05 Thread Martin Pfeifle
Thanks, If we look at a B+-tree, all records are stored at the leaf level and form at least a logical list. But what happens, if we insert new records which do not fit on the corresponding leaf page any more. Assume this page has to be split. Where is the newly created physical page stored? Does

AW: [sqlite] primary key and physical data organization

2006-04-05 Thread Martin Pfeifle
That's great thank you very much. - Ursprüngliche Mail Von: [EMAIL PROTECTED] An: sqlite-users@sqlite.org; Martin Pfeifle <[EMAIL PROTECTED]> Gesendet: Mittwoch, den 5. April 2006, 23:09:25 Uhr Betreff: Re: [sqlite] primary key and physical data organization Martin Pfeifle <[EMAIL PROTE