Re: recursive relationship (having trouble posting to this newserver, so excuse repetition)

Yep, I think I've seen this reply about 6 times over the last few days. You can probably stop resending it now.

reply-to_at_cec-services.com wrote in message <85lnrt$t5o$3_at_dosa.alt.net>...
>From Fernando <spamers_at_must.die>:
>
>> On Wed, 12 Jan 2000 21:50:00 GMT, joe_celko_at_my-deja.com wrote:
>>
>> >>> I do not understand very well the concept of a recursive
>> >relationship. Could someone please explain and give me an example? <<
>>
>> [snip]
>>
>> Thanks Joe, it was much more than I expected. :-)
>
>Actually Joe could be mistaken because a recursive relationship is possible
>where SQL code in consecutive triggers effectively calls the same tables.
>
>This is used in real time manufacturing, among other places, and maps
 perfectly
>to the reduced Petri Net form for a just-in-time Kanban cell.
>
>The pseudo code is:
>
>let column.N = column name indexed by the value of N
>let value.T = a test switch value of interest of T
>
>select rows.A
>from table.A
>where column.N = value.T
>
>select rows.B
>from table.B
>where column.B = rows.A [returned from above]
>
>select rows.A
>from table.B
>where column.A = rows.B [returned from above]
>
>or in the format of one trigger:
>
>select rows.A
>from table.B
>where column.A = (select rows.B
> from table.B
> where column.B = (select rows.A
> from table.A
> where column.N = value.T))
>
>When implemented and tested, the performance of the above pseudo code is as
>follows on three RDBMS vendors in arbitrary units of the same time:
>
>IBM DB2 100
>ORACLE 600-1200
>MS SQLS 1200-2400
>
>Therefore ORACLE is 2 times faster than MS SQLServer, IBM DB2 is 6 to 12
 times
>faster than ORACLE, IBM DB2 is 12 to 24 times faster than MS SQLServer.
>
>Performance of the RDBMS engine becomes critical in real time processing
 such as
>for incoming call data in billing systems for cellular communications
 companies.
>
>The above effectively brings real time performance to RDBMS technology by
>coercing non procedural SQL to perform procedural processing.
>
>The development and maintenance advantage is that SQL code is no longer
 embedded
>in multiple non portable high order languages but is 100% portable across
>platforms in only one set of code.
Received on Fri Jan 14 2000 - 00:00:00 CET