Hi All, I think a local development is an alternative, since a history mostly use for reporting purpose a live replicated template will do the trick, if you want only the last change of the year than id for the record will like "TF08XXXXX.YYYY" where YYYY represent year like 2007, 2008 so on. if you want to record all the changes in a year you can add CURR.NO in the id too than your record id will be TF08XXXXX.YYYY.ZZ where ZZ is curr.no obviously you need to write a batch routine, some other auth routines for replication, Regards Ahmed Ali
On 9/20/09, Jim Idle <[email protected]> wrote: > > On 09/19/2009 10:53 AM, JOSE L MARTINEZ-AVIAL wrote: > > I recall doing that long time ago. The problem was that you need to create > a routine that receives an ID, and returns the number of the field in the > distribution (Each distributed file is associated to a number) That > algorithm needs to be constant(I don't remember the correct name) in the > sense that for a certain ID the algorithm should return always the same > number. Otherwise it wouldn't work. For example, if today (2009) you > modified a record that was created in 2008, the ID (which is the only > information you can use in the routine) would be something like TF08XXXXX. > And it should be written in the 2008 file(You can not use the current date > as a key in the routine, because in 2010 you will be unable to find the > record). So what you are asking doesn't seem to be possible. > > > You get the record too and presumably that has a date field which means you > can calculate the correct partition. Trivial. Just look up the partioned > file documentation. Also, remember that you can changed the ID that is used > to store the record in the part file to influence the hashing order (and get > significantly better performance for selects and so on sometimes). You need > to be able to calculate the new ID as the same every time and calculate teh > original id from the stored (in the partition file ID).\ > > In other words both the partition you assign and the ID you assign (if you > change it) must be deterministic, and not transient. > > Jim > > > Regards > JL > 2009/9/18 andi <[email protected]> > >> >> We need to distribute some T24 History files based on the Time. Dose >> anyone has a suggestion. >> Ex: FBNK.DRAWINGS$HIS >> We need to Store the Distributed data Yearly. >> >> i.e. >> History Data of year 2006 should be in FBNK.DRAWINGS.HIS.2006 >> History Data of year 2007 should be in FBNK.DRAWINGS.HIS.2007 >> History Data of year 2008 should be in FBNK.DRAWINGS.HIS.2008 >> History Data of year 2009 should be in FBNK.DRAWINGS.HIS.2009 >> >> BASIC REQUIREMENT IS THAT NEW HISTORY RECORDS SHOULD NOT UPDATE >> PREVIOUS YEARS FILES. >> >> >> > > > > > > > > -- Ahmed Ali Nisar --~--~---------~--~----~------------~-------~--~----~ Please read the posting guidelines at: http://groups.google.com/group/jBASE/web/Posting%20Guidelines IMPORTANT: Type T24: at the start of the subject line for questions specific to Globus/T24 To post, send email to [email protected] To unsubscribe, send email to [email protected] For more options, visit this group at http://groups.google.com/group/jBASE?hl=en -~----------~----~----~----~------~----~------~--~---
