We have a bunch of data that users can view in windows and do it quickly. They may want to look at the data window, which is a day, week, month, or arbitrary start and end data. Sorting and summing all this material in real time is painful for us, so I got the idea to do something similar to Mipmaps in 3D rendering. Ultimately, you save the same data, previously calculated at different scales, and then interpolate the results using different scales. Therefore, I would already know what numbers were during the year, a certain month, a given week, and a certain day for the store, and if they ask for a certain range, I use different scales to quickly add something,which gives me the right, but I don’t need to process the complete data set, I just extract four or five records and add or subtract them.
Is this a real model? Does it make any sense and there are places that I can read about how to do this best, or are there much better ways to deal with large chunks of data, for example, where they need to be viewed in different slices?
It seems that this should be a well-known and resolved problem. For example, many people have stock portfolios, and every day they have to do such things. Our data are not stock prices, but the idea is the same.
source
share