On Dec 14, 2008, at 2:45 PM, robert rottermann wrote:
> > Hi there, > > I have a query like: > > sesion.query(MytableTable) > > which returns a huge numer of recors. > now I would like to iterate trough the records. > What method should I use? > > sesion.query(MytableTable).all() returns all > and next or fetchone is not defined. > > thanks for your help > robert The Query is already an iterator and can be used as such. However it prefetches the full result by default, since the ORM must complete its bookkeeping on any ORM-mapped object instance before returning it, and it's very common that a single instance will appear multiple times in a result set - such as if the query is joining multiple tables together, multiple rows which contain the same primary key columns will be translated to a single object instance in the identity map. If you'd like to iterate through a very large result set without memory growing to the size of the result, you can use query.yield_per(<some number of rows>) and iterate that - "some number of rows" will be prefetched at a time. When using this option, you will have to ensure that either A. the query returns only individual columns and not ORM mapped objects or B. the query's statement is guaranteed to represent each returned ORM instance uniquely - once an ORM mapped instance is yielded, if the Query encounters it again it will process it as though it were new and may cause undefined behavior. As an alternative, use LIMIT/OFFSET options to limit the size of the result set represented by a single statement execution/result iteration. --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the Google Groups "sqlalchemy" group. To post to this group, send email to sqlalchemy@googlegroups.com To unsubscribe from this group, send email to sqlalchemy+unsubscr...@googlegroups.com For more options, visit this group at http://groups.google.com/group/sqlalchemy?hl=en -~----------~----~----~----~------~----~------~--~---
