Actually, I also considered that. The reason I choose to use a boxed structure is for support. My production code (really different from the sample but similar concept) has data coming in from SQL server. I technically have "named" columns. So I implemented an inverted table where each column is stored as a list in a Named variable (the actual database field name, ie. FieldName1, FieldName2, FieldName3). XML_USER_MAP +----------+----------+----------+ | <name> |FieldName1|</name> | +----------+----------+----------+ | <city> |FieldName2|</city> | +----------+----------+----------+ | <country>|FieldName3|</country>| +----------+----------+----------+
This is crucial in the long run because the MAP's are actually in a separate text file where somebody (a developer hopefully) can edit the text file and put in the appropriate database table field name into the appropriate tags. Essentially, if they change the database structure, all they need to do is change the values in the map text file and doesn't have to bother me about it. :P -----Original Message----- From: [email protected] [mailto:[email protected]] On Behalf Of Raul Miller Sent: Wednesday, May 05, 2010 4:45 PM To: Programming forum Subject: Re: [Jprogramming] Poor mans way of creating XML document files in J Personally, I would be inclined to build the whole thing based on rplc. -- Raul ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
