Re: AW: DDL Discussion continued

A DDS defined logical file can have features similar to an SQL index and a view.

IBM simply chose to have them show up under view instead of index in
iNav. Given that most DDS LF's I've seen have keys and the same
format as the physical, I would have put them under indexes. I can't
say that I've ever seen an non-keyed DDS LF.

You can DBU/DFU or use RPG and access an SQL defined index just like a
keyed DDS defined logical.

Thus, it's easy to replace most DDS defined logicals with an SQL
index, and there are performance benefits from doing so:
http://www-03.ibm.com/systems/resources/systems_i_software_db2_pdf_Performance_DDS_SQL.pdf

The SQL defined indexes default to a larger page size simply because
the lager size makes sense now-a-days given the amount of RAM common
in today's boxes. IBM chose to not to change the size of DDS LF,
perhaps the smaller size makes sense for RLA.

However. If you create a SQL index, with keys KEYFLD1, KEYFLD2,
KEYFLD3 and then create a DDS logical with the same keys. The logical
file will share the access path contained in the SQL index! DB2 won't
bother to create a duplicate access path in the DDS logical object.

HTH,
Charles





On Wed, Oct 1, 2008 at 2:45 PM, Dave Odom <Dave.Odom@xxxxxxxxxxxx> wrote:

Vern, et al,

Where I get my info there is a difference between a LF that may be an access path and an true index created by the CREATE INDEX command, is from how they look in the system and via Navigator. I"m also not sure a LF is constructed underneath like an index either but you all probably know more about that.

When I take a look at a logical file it seems to have a structure, much like a PF. You can see actual data. When I take a look at an index, I don't see the same thing. The LF also doesn't seem to have the same page size, the index being much bigger. It seems in a LF, you can have SQL to create a view and therefore do things that don't seem to be available with an index. It appears a LF can be accessed via a program to sort and present the data to a program in a different way than how the data exists in a PF. An index is accessed "under the covers" and from what I can see, can't be accessed directly by a program. Just some of the differences I see.

On the system I work on, there a MANY LFs/access paths but they don't show up as indexes via Navigator but an index does.

That, coupled with the traditional use of an index in a relational database vs. a VIEW (LF to you all) makes them different animals. Why are my observations so different between the two but yet you all say they are the same? What am I missing?

Dave


--
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list
To post a message email: MIDRANGE-L@xxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: http://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxx
Before posting, please take a moment to review the archives
at http://archive.midrange.com/midrange-l.