If your main concern is to store scientific data on disk you might try: http://www.pytables.org/moin
However, it uses numarray internally and a C library, which you have to build from source. (You use a Mac right?) Concerning your code: - Your two file solution seems impractical to me. I think you should just pickle your whole dbase object. - Maybe you should write 'load' and 'store' methods that create the temporary file, Pickler and Unpickler objects. -The __init__ method should then construct the object from a list of variable names and an array. -Offcourse you need a set method. more ideas: - A special variable name 'time'. Then you can implement a getAtTime( varNameList, timePoint) method with interpolation. - A 'plot' method that works like matplotlib's plot function. - An extract(varNameList) method, that returns a new dbase object with only the selected variables. - A companion class that can hold several time series at once to compare different experiments. Finally, post the code to the mailing list. At least I would like to use such a class :-). Yours Eike. _______________________________________________ Numpy-discussion mailing list [email protected] http://projects.scipy.org/mailman/listinfo/numpy-discussion
