Op vrijdag 10-10-2008 om 20:31 uur [tijdzone +0200], schreef Dusan Halicky: > I'm trying to create "lightweight xml dataset" derived from TDataSet. > I've studied various other datasets (TFixedFormatDataset, > TParadoxDataset, TMemDataSet) but it's too complicated, perhaps > because there are no comments at all. Can someone help me at least > with some info how it works? Is there any documentation on datasets?
Yup, it ain't easy. But I think that TBufDataset is the best one to look at. What you could also do is use TBufDataset as parent for your dataset. That way you only have to implement a few methods. > Right now I'm trying to create some dataset which return static data > (e.g. table of numbers from 1 to 100 in column1 and random strings in > column2, some boolean in column3). I finally have something in the > dbgrid but it acts weird, I have no idea how to handle data. You probably need to implement RecNo, or bookmarks. > The reason I need my own dataset is that I am creating 3-tier > application in lazarus. The key feature why I need proper dataset is > that If I select too many data, e.g. 10k rows table, the dataset will > fetch only few lines and then when user scroll the dbgrid new data > will be fetched. As far as I know, this can be done only in TDataSet. That's all implemented in TBufDataset. It you create a descendent of TBufDataset, you're done. Joost. _______________________________________________ fpc-devel maillist - fpc-devel@lists.freepascal.org http://lists.freepascal.org/mailman/listinfo/fpc-devel
