[Date Index] [Thread Index] [Author Index]
Re: How can a dump file with 1.28MB have nothing in it?
David, et al., I think the answer to my own question is that even an empty dump file will contain >1MB. What probably happened was that I attempted to save to the dump file from a notebook with an empty or no kernel. I use dump files because I'm doing what might be called exploratory data analysis. The methods I developed seem to be like no others in the statistics and chemometrics packages I know of. Since these methods are new, they can't be encapsulated in a program. Perhaps a package of functions would help somewhat, but I still have long, experimental sessions in which the data are transformed and manipulated in all manner of ways to find an acceptable problem solution. The problem is not so much procedures that number crunch for a long time, although that happens sometimes, but an evolving process in which it is most convenient to keep intermediate results of all sorts available in the kernel for examination and re-use. Hence the use of Dump files to store a snapshot of data and variables in process. As long as I'm working this way, I need to be more disciplined about saving final results and frequently "backing up" via DumpSave. Also, I am now DumpSaving to 2 different file names per session. If you have any ideas for a better strategy to do my kinds of problems, I would be glad to hear them. Thanks for the response so far. On Dec 25 2006, 2:07 am, David Bailey <dave at Remove_Thisdbailey.co.uk> wrote: > doesitinmud wrote: > > I have been saving my intermediate results to adumpfileduring the > > course of a very large notebook (~10MB, ~800 lines). I decided to > > begin a new notebook, and begin it by opening thedumpfile. When I > > started a new notebook and loaded thedumpfile, I could no longer find > > any of the symbol definitions saved from the old notebook. How did > > this happen? How can I retrieve the information in thedumpfile?How can anyone answer this question - as you have supplied no code! I > must say, however, I would not store important results in binary format > - it probably takes up less disk space, but if something goes wrong, you > have nothing. If you ARE going to use DumpSave, I would at least check > that it is working as you expect early on in the process. > > Are you performing a calculation that takes an excessively long time to > run, and if so, are you sure this could not be speeded up in some way so > that there is no need to do the calculation in several stages. > > If you supply more details of your problem, I am sure someone here will > be able to help - although probably not to salvage your previous results! > > David Baileyhttp://www.dbaileyconsultancy.co.uk