Thread: hello
Hello, I just joined the pgAdmin mailing list. I would like to say that I am generally very impressed with the latest version of this project. The UI is much tighter than version two and the whole thing feels much more usable. One thing that I have noticed is that the 'view data' tool has problems with large (> 50k) result sets. I am very familiar with this problem and have dealt with it myself in various projects. IIRC a cursor based solution was marked 'controversial' in the TODO list...is this still the case? (I'm grabbing the latest sources right now). I've been wanting to learn the wxWindows toolkit for a while now and this seems a good place to start. :) Merlin
> -----Original Message----- > From: Merlin Moncure [mailto:merlin.moncure@rcsonline.com] > Sent: 10 May 2004 18:33 > To: pgadmin-hackers@postgresql.org > Subject: [pgadmin-hackers] hello > > Hello, > > I just joined the pgAdmin mailing list. I would like to say > that I am generally very impressed with the latest version of > this project. The UI is much tighter than version two and > the whole thing feels much more usable. > > One thing that I have noticed is that the 'view data' tool > has problems with large (> 50k) result sets. I am very > familiar with this problem and have dealt with it myself in > various projects. IIRC a cursor based solution was marked > 'controversial' in the TODO list...is this still the case? > (I'm grabbing the latest sources right now). I've been > wanting to learn the wxWindows toolkit for a while now and > this seems a good place to start. :) Hi Merlin, good to have you aboard :-) Yes, the cursor based *query tool* is marked as controversial. Andreas & I have always had a differing of opinion over what it's true purpose is (and I don't think either of us are actually wrong if you know what I mean). Iirc, Andreas' main objection to using cursors is that he likes to use the query tool to refine queries, and if it's using cursors he can't properly assess the speed gains of one query over another. That argument doesn't apply to the view data tool though - I see no reason why that cannot be made cursor based. If you're working on it anyway, the ability to copy/paste 1 or more rows both back into the table and into external spreadsheets etc. would be nice!! <grin> Regards, Dave.
Dave Page wrote: > > > > >>-----Original Message----- >>From: Merlin Moncure [mailto:merlin.moncure@rcsonline.com] >>Sent: 10 May 2004 18:33 >>To: pgadmin-hackers@postgresql.org >>Subject: [pgadmin-hackers] hello >> >>Hello, >> >>I just joined the pgAdmin mailing list. I would like to say >>that I am generally very impressed with the latest version of >>this project. The UI is much tighter than version two and >>the whole thing feels much more usable. >> >>One thing that I have noticed is that the 'view data' tool >>has problems with large (> 50k) result sets. >> I don't know about any problems with large result sets. Of course there's some delay until the server has done selecting the whole table (or the filtered subset), but since the data is *not* transferred until needed there's no space or data delivery problem. Not to speak of the *human* problem handling such large data sets. >> I am very >>familiar with this problem and have dealt with it myself in >>various projects. IIRC a cursor based solution was marked >>'controversial' in the TODO list...is this still the case? >>(I'm grabbing the latest sources right now). I've been >>wanting to learn the wxWindows toolkit for a while now and >>this seems a good place to start. :) >> >> > >Hi Merlin, good to have you aboard :-) > >Yes, the cursor based *query tool* is marked as controversial. Andreas & >I have always had a differing of opinion over what it's true purpose is >(and I don't think either of us are actually wrong if you know what I >mean). Iirc, Andreas' main objection to using cursors is that he likes >to use the query tool to refine queries, and if it's using cursors he >can't properly assess the speed gains of one query over another. > >That argument doesn't apply to the view data tool though - I see no >reason why that cannot be made cursor based. If you're working on it >anyway, the ability to copy/paste 1 or more rows both back into the >table and into external spreadsheets etc. would be nice!! <grin> > > Yes, copy/paste of rows from view data (and query tool as well) would be helpful and is one of the most often requested features. Regards, Andreas
> Yes, copy/paste of rows from view data (and query tool as well) would be > helpful and is one of the most often requested features. Questions about wx toolkit: Does wxString have fast repeated concatenation for large strings (like STL string)? Does wxGrid have the capability to demand load row contents from an externally managed source? What format? Usually in these situations I tend to use tab delimited for easy pasting into exc--er, open office :) I took a quick look at the code and it seems very straightforward. Maybe I'll have a crack at this... WRT the query tool, I mostly use it like Andreas does. But I also agree that there is not one universal data browser that can meet all the various requirements. What has me concerned is that pgAdmin lacks the capability to provide editing on large tables, because of memory issues with the editor and the lack of editing in the query tool. I think this is a solvable problem :) with the ability to provide editing over infinite sets I think pgAdmin will become a first class data manager (or more so than it already is) :). Merlin
> >Questions about wx toolkit: > >Does wxString have fast repeated concatenation for large strings (like > >STL string)? > > > > > > > Somehow, don't know if it's enough for your purpose. Would any code using STL containers be accepted into the project? Merlin
Merlin Moncure wrote: That's an awful lot of questions on wx specific stuff, you'd better go to wxdev for such things. >>Yes, copy/paste of rows from view data (and query tool as well) would >> >> >be > > >>helpful and is one of the most often requested features. >> >> > >Questions about wx toolkit: >Does wxString have fast repeated concatenation for large strings (like >STL string)? > > > Somehow, don't know if it's enough for your purpose. >Does wxGrid have the capability to demand load row contents from an >externally managed source? > > > pgadmin3 does so in view data, overload wxGridTable >What format? Usually in these situations I tend to use tab delimited >for easy pasting into exc--er, open office :) > > For the start, use the export options (I personally never use tabs but | or ;). >I took a quick look at the code and it seems very straightforward. > > I did my best. >Maybe I'll have a crack at this... > > >WRT the query tool, I mostly use it like Andreas does. But I also agree >that there is not one universal data browser that can meet all the >various requirements. What has me concerned is that pgAdmin lacks the >capability to provide editing on large tables, because of memory issues >with the editor > Definitely not! >and the lack of editing in the query tool. I think this >is a solvable problem :) > No, this is how pgadmin2 worked, and it's a mess. Use view data with filter, it's much cleaner. > with the ability to provide editing over >infinite sets I think pgAdmin will become a first class data manager (or >more so than it already is) :). > > > :-) Regards, Andreas
Merlin Moncure wrote: >>>Questions about wx toolkit: >>>Does wxString have fast repeated concatenation for large strings >>> >>> >(like > > >>>STL string)? >>> >>> >>> >>> >>> >>Somehow, don't know if it's enough for your purpose. >> >> > >Would any code using STL containers be accepted into the project? > > What should this be good for? This probably affects portability, so you must have a *very* good reason to do so. Regards, Andreas
> What should this be good for? This probably affects portability, so you > must have a *very* good reason to do so. Understood :)...never mind. I like using STL strings and string builders because they are very fast, likewise for iterators and sorting. Nothing I can't do without, though. Merlin
Merlin Moncure wrote: >Understood :)...never mind. I like using STL strings and string >builders because they are very fast, likewise for iterators and sorting. >Nothing I can't do without, though. > > Please use the wx equivalents, which are going to be merged with stl versions anyway in the far future (probably as a compile option). Regards, Andreas