Version: (using KDE 3.5.8) KSpread (comparing to OOffice and "common sense") is much too slow while creating the preview (in OOffice it is instantaneous), and loading the file is pain. This is so odd that I attach the file I am using. Related report on speed: https://bugs.kde.org/show_bug.cgi?id=59510
Created attachment 23715 [details] test.csv.bz2
hi Maciej, Bug #59510 is solved in trunk aka upcoming KSpread 2.0. But this report is imho not related since the difference between OOCalc and KSpread here is, that KSpread always loads the whole cvsfile for preview. iirc e.g. Kexi, that adopted the cvs import code (see also Jaroslaw's comment at bug #158462 here), did fixed it already by limiting the previewed lines to ~100 and just ignore every additional line what makes sense since it's a preview only anyway. So, if KSpread would do it the same way, we wouldn't be able to provide a progressbar what renders bug #158462 somewhat useless except there is a clever way to use readed lines vs. total filesize ;)
Sebastian, if KSpread loads the whole file for preview then what takes so long to load it fully? It should be a blink in such case. The progress bar is nice, but without it I will live :-) Maybe I explain -- I just like some kind of feedback. Currently Kspread freezes (GUI I mean) while loading files -- which is bad, because when you wait you cannot be sure, if program is already hanged up for good or everything is ok, and data will be loaded finally. Anyway -- speedup is very welcome :-))
Maciej, cause the attached files has ~25.000 lines and each of that line needs to be parsed. There is a huge difference if you do this for 100 lines or for 25k lines in a preview and it's not like in a simple texteditor that does not need to go through each single line, split it, try to determinate the type (numeric, string, date, etc.) and whatever else.