Vineet, Agreed, the temp file should be in the DAL. I don't see at this stage why it would be break the logic.
On May 17, 5:47 am, Vineet <vineet.deod...@gmail.com> wrote: > Simon, > I thought on the idea of adding a MySQL temp table to manage insert/ > del/update on multiple tables. > But there would be a large no. of such forms requiring this > functionality. > It is better to avoid fiddling with the schema (which is the database- > tier) for the sake of business-logic tier. > > What if I add a temporary table in DAL itself (along with the other > linked tables)? > Would it break the web2py logic of generating, validating, and saving > the form (if this table is used for generating the form)? > I mean, if a form is generated from such a temp DAL table, whether it > is possible to save data in multiple different tables. > > ---Vineet > ============================================ > > On May 16, 4:33 pm, Simon Ashley <gregs...@gmail.com> wrote: > > > > > > > > > Memory tables referred to have no ability to write to disk. Generally > > we've used these for report accumulations. You could use a MySQL table > > and delete the record at the end (if need be). Crude and there should > > be a more pythonic method, but should work. > > > Define temp table fields as per the following: > > > 1. rec id > > 2. table1.uid > > 3. table1.variable > > 4. table2.uid > > 5. table2.variable > > 6. table3.uid > > 7. table3.variable > > > Populate the fields before the load. The uids will of course be hidden > > on the form but used after the submit in separate database actions. > > > Zero uid's would trigger inserts, > 0 for deletions where variables > > are blank or zero, edits where not. There is a level of programming > > involved; Massimo implied that. > > > In your case, I would be reluctant to denormalise as further > > complexity could be cumbersome. > > (if not unworkable. Can think of examples in an application that we > > are looking at converting that would make anyone's mind spin). > > > YMMV > > > On May 16, 1:15 am, Vineet <vineet.deod...@gmail.com> wrote: > > > > Simon, > > > That is a good idea. > > > In fact, almost the same approach is being followed in VFP cursor > > > adaptor (the difference being that, data from MySQL tables is > > > reproduced in same table structures, i.e cursors). > > > > Here, by virtual table, do you mean to say that we should reproduce > > > the data into dictionaries by fetching a dict cursor from MySQLdb? > > > (in DABO, they have an excellent concept of bizobj for this task). > > > > I am not asking for exact code, but can you suggest an approach to > > > tackle the task? > > > > Dear Massimo & other friends, > > > can you pl. suggest something to solve my problem? > > > > Thanks, > > > Vineet. > > > > ================================================================= > > > On May 15, 5:07 pm, Simon Ashley <gregs...@gmail.com> wrote: > > > > > Since no one has replied for a couple of days, I'll toss my 2 cents > > > > in. Don't how much value this could be, as we just evaluating web2py > > > > and python but in another environment we could have written data in > > > > the normalised records to a temporary record in a denormalised, > > > > virtualtable, update and saved this record, and then do selective > > > > updates, insertions and deletions on each normalised record in the > > > > original tables. Not that elegant but ... > > > > > On May 13, 9:17 pm, Vineet <vineet.deod...@gmail.com> wrote: > > > > > > pbreit, > > > > > As you say, wherever possible or required, I have already denormalized > > > > > thetable-structures. > > > > > Apart from normalization / denormalization issue, in many other forms, > > > > > I need to insert/update/delete records inmultipletables from one > > > > > form. > > > > > (as I mentioned earlier, there will be a mix of insert/update/delete > > > > > of record(s) in different tables.) > > > > > > Massimo, > > > > > If you too haven't understood what I intend to do with these 5 tables, > > > > > I can elaborate the idea. > > > > > Pl. tell me accordingly. > > > > > The very success of developing my project using web2py hinges on this > > > > > functionality. > > > > > > ---Vineet > > > > > ================================================ > > > > > > On May 11, 5:17 am, pbreit <pbreitenb...@gmail.com> wrote: > > > > > > > Normalizing is not always the best approach. If the database has > > > > > > not yet > > > > > > been designed, perhaps reconsider. I still don't really understand > > > > > > what you > > > > > > are trying to do but 5 small tables might not be necessary. > > > > > > > I would suggest getting some code working with 2 tables an then go > > > > > > from > > > > > > there. > > > > > > > Some > > > > > > options:http://web2py.com/book/default/chapter/07#One-form-for-multiple-table... > > > > > > > Another approach would to use FORM or SQLFORM.factory and then > > > > > > coding the > > > > > > database updates. You lose some of the Web2py features but it might > > > > > > still be > > > > > > the easiest way to do it. > > > > > > > But my first suggestion would be to try and simplify what you are > > > > > > intending > > > > > > to do.
