Reloading data to a sorted set

Discussion created by ca.portal.admin on Aug 5, 2005
Latest reply on Aug 5, 2005 by ca.portal.admin
Hello List,

I have a question dealing with the mass loading of records.

We will be extracting all of the member records from a sorted set.
seq. DSC by store date duplicates first) After our Database group
initializes and resizes our area. We will then be re-storing a portion
the member records. We are trying to figure out the most efficient way
reload the data.


We are currently planning to unload the member records oldest to newest
each of the owner records. One idea we had was to ask our DBA staff to
create a temporary subschema for the load of these records so that the
member record would be stored first as shown below. Thinking that IDMS
would have to do less work if it didn't have to maintain the sorted set.


So, our first question is does this idea have any merit? Secondly we
ask if this is not the best way to go about unloading and reloading this
data while maintaining the sequence of the records how should we be
attacking it?


Scott Van Der Heyden
State of Iowa
Dept. Of Human Services

IDMS Public Discussion Forum


Re: Reloading data to a sorted set
what about the duplicates first constraint? If that order still has
to be kept, an additional sort key would be necessary. The rest of the
idea is good assuming there are no cross area sets into the target area.
Note also that calc recs should be loaded in descending order of
suggested page # in order to minimize calc overflow.

I would suggest not creating a special unsorted load subschema. If
the member records will be sorted for the reload, there will be no
additional pointer update work for IDMS to perform and you'll know for
certain when all is done that the order is the way you want it.
According to the DBA book, the storage order of the member records
should mimic the intended sort order, in other words, newest to oldest.