Re: Oracle read consistency *inside*a transaction

From: VC <boston103_at_hotmail.com>
Date: Wed, 08 Dec 2004 00:13:47 GMT
Message-ID: <_Grtd.157418$5K2.20101@attbi_s03>

Please see below:

[...skipped for brevity ...]

[AF]

>> > "Timestamps" for uncommitted transaction, as well as any other
>> > transactional information, is stored in internal transaction tables
>> > (that's why I referred you to Concepts Manual, it's all described
>> > there).
>>

[VC]

>> 1.  What specific "internal transaction table" stores information

> about

>> the "timestamps" for uncommitted transactions ?
>>
>> 2.  Assuming there  is an internal table with the "timestamps",  how
>> exactly does Oracle use the "timestamps" to solve the problem I've
>> described in my original message ?
>>
>> 3.  What page of the Concepts,  say version 10g,  decribes the table
>> along with the "timestamps" and steps Oracle takes to solve the

> problem

>> I described earlier ?
>>
>> Thanks.

[AF]

> Internal workings of transaction management is not published by Oracle.
> You can't access internal transaction tables, Oracle does not publish
> algorithms it uses for transaction management.
> Concepts Manual references transaction tables here:
>
> http://download-west.oracle.com/docs/cd/B14117_01/server.101/b10743/transact.htm#sthref613

[VC]

Well, well... The document sez:

[quote]
"When a transaction is committed, the following occurs:

The internal transaction table for the associated undo tablespace that the transaction has committed, and the corresponding unique system change number (SCN) of the transaction is assigned and recorded in the table." [/quote]

Apparently, Oracle cannot afford hiring an English-speaking techical writer or a proof-reader any longer because this:

" The internal transaction table for the associated undo tablespace that the transaction has committed,"

... does not make any sense whatsoever.

Yet, assuming the "internal transaction table" hidden from hoi polloi does contain a "timestamp" and further assuming from the cryptic phrase above that "the unique system change number (SCN)" is such timestamp, my original question remains unanswered: how does Oracle distinguish between two changes inside the same transaction ? Clearly, the SCN cannot be used for this purpose because both changes would have the same SCN/timestamp.

[AF]

> and here:

> http://download-west.oracle.com/docs/cd/B14117_01/server.101/b10743/process.htm#sthref1576
>

[VC] Sorry, the referenced page does not say a word about "internal transaction tables".

Any other thoughts ?

Thanks. Received on Tue Dec 07 2004 - 18:13:47 CST