My application needs to get data for a chart from either a file on the local computer or via http. I'm trying to work out what will be the BetterSolution(TM) : reading in all of the data into a dataset at the start, or implementing some functions to pull data from the file as it is needed. As far as I can tell the latter won't really be possible for http access as they would need to be able to seek to arbitrary points in the file; but this should still be possible for local files. The former idea would be inefficient for large datasets.
any thoughts/experiences?
vik
datasets from files/URLs
-
vik
Re: datasets from files/URLs
Not sure yet 
But I'm planning for the worst. I'm graphing data from a datalogging application, so it really depends how far back the user would want to go. I'm trying to decide if the data should be broken into segments, say of one month or one week. If data was sampled every 5 minutes, this would give about 2000 samples per week. So I'm looking at between 10^3 and 10^4 data points per series, with perhaps 5 series.
cheers
vik
But I'm planning for the worst. I'm graphing data from a datalogging application, so it really depends how far back the user would want to go. I'm trying to decide if the data should be broken into segments, say of one month or one week. If data was sampled every 5 minutes, this would give about 2000 samples per week. So I'm looking at between 10^3 and 10^4 data points per series, with perhaps 5 series. cheers
vik