Hi Joe, Depending on how many different structures you have you may want to consider having some external configuration object. My first thought when reading your description was that there are only so many ways to parse, and so many different types of data. Having a collection of parser objects that handle specific translations seems clean. Once you have an idea of what types of translations you need you could trigger those operations by building your collection of parser objects from an external config file. In that way you read the parameters in (could be a file in the JPMorgan directory) conf.xml or parse.conf something like that, and then use that to set up your parser. The parser is then written to be generic and reusable. It would allow you to write methods once and then reuse them for files you haven't seen yet by creating a new config file for your format. You could even have a config file generator that asks you questions and shows you the results from a current sample data file. :)
fieldSeperator: #comma. fieldDelimited: #doubleQuote. nameSeperator: #comma. nameFormat: 'title, first, [mi], last, [suffix]'. balance: #USD. fieldOrder: 'id, name, balance'. ... One other thought is if there is a way for the system to determine what format to use (this is a JPMorgan file) then instead of an external config file you could just store the different configs internally and match for the right config collection and error if one doesn't exist (asking the user to create one using your config builder method). if you can't match then having a file in a JPMorgan directory seems simple enough. In general thinking of the setup step as a collection of generic parser configuration objects instead of a different parsing method for each file, simplifies everything. Of course back to my original point, writing 3 parser methods will go faster if there are not that many formats. There is always a trade off when you consider building a framework or just hacking some code that works :). All the best, Ron Teitelbaum > From: Joseph Alotta > Sent: Thursday, April 28, 2016 6:15 PM > > Greetings, > > I am writing a program to consolidate all my personal finances for tax time > next year. (This is not a school project.) > > There are transaction files from several banks and credit card companies. > Each has a similar format, CSV, but they vary in many ways, order of items, > extra notes, pipe delimited or tabs, etc. I want to read them and load them > into a collection of transaction objects. > > 1. Should I have a FileReader object? > > 2. Should it have subclasses like FileReaderAmericanExpress, > FileReaderJPMorgan ? > > 3. Or should it have different methods like loadAmericanExpresFile, > loadJPMorganFile ? > > 4. Is a Collection of Transaction objects, the structure that you would load > the files into? > > The rest of the project would be to do data checking on the files, to make > sure there are no duplicates or missing dates. Then write reports that I can > give to my accountant. > > I would appreciate some design help? > > Sincerely, > > Joe._______________________________________________ > Beginners mailing list > [email protected] > http://lists.squeakfoundation.org/mailman/listinfo/beginners _______________________________________________ Beginners mailing list [email protected] http://lists.squeakfoundation.org/mailman/listinfo/beginners
