Re: parallel query

In article <3hd4b0$arj_at_dcsun4.us.oracle.com>, edansker_at_us.oracle.com (Edward L Dansker ) writes:
>Paul Singer <singerap_at_powergrid.electriciti.com> writes:
>
>>alal1_at_PROBLEM_WITH_INEWS_GATEWAY_FILE (Anuj Lal) wrote:
>>>
>>>
>>>
>>> I heard oracle 7.1 has parallel query option.
>>> I have single query which has process table with 10 million rows
>>> and fetch around 50000 rows.
>>> How can I use the paralle query option for this query.
>>> Do i have to declare different cursor for each process fetching data
>>> from same table.
>>> Thanks
>>> alal1_at_eccdb1.pms.ford.com
>>>
>>
>>
>>There are 2 ways that this works. The first is that you alter the
>>table and declair access to it to be parallel. The syntax is
>>specified in the 7.1 documentation ademdum that should have been
>>shipped with your 7.1 version.
>>
>>The second way is to embed a hint in your query something like
>>this:
>>
>>SELECT --+ PARALLEL(table alies, degree of parallelism, number
>> of nodes )
>> * FROM table ...
>>
>>Again, the syntax is sepecifed in the documentation addendum.
>
>Please note that this will only be beneficial if you have multiple processors
>that you can dedicate to this task. A single-CPU machine cannot take advantage
>of parallelism.
>
>Ed
>
 In article <3hd4b0$arj_at_dcsun4.us.oracle.com>, edansker_at_us.oracle.com (Edward L Dansker ) writes:
>Paul Singer <singerap_at_powergrid.electriciti.com> writes:
>
>>alal1_at_PROBLEM_WITH_INEWS_GATEWAY_FILE (Anuj Lal) wrote:
>>>
>>>
>>>
>>> I heard oracle 7.1 has parallel query option.
>>> I have single query which has process table with 10 million rows
>>> and fetch around 50000 rows.
>>> How can I use the paralle query option for this query.
>>> Do i have to declare different cursor for each process fetching data
>>> from same table.
>>> Thanks
>>> alal1_at_eccdb1.pms.ford.com
>>>
>>
>>
>>There are 2 ways that this works. The first is that you alter the
>>table and declair access to it to be parallel. The syntax is
>>specified in the 7.1 documentation ademdum that should have been
>>shipped with your 7.1 version.
>>
>>The second way is to embed a hint in your query something like
>>this:
>>
>>SELECT --+ PARALLEL(table alies, degree of parallelism, number
>> of nodes )
>> * FROM table ...
>>
>>Again, the syntax is sepecifed in the documentation addendum.
>
>Please note that this will only be beneficial if you have multiple processors
>that you can dedicate to this task. A single-CPU machine cannot take advantage
>of parallelism.
>
>Ed
>
 In article <3hd4b0$arj_at_dcsun4.us.oracle.com>, edansker_at_us.oracle.com (Edward L Dansker ) writes:
>Paul Singer <singerap_at_powergrid.electriciti.com> writes:
>
>>alal1_at_PROBLEM_WITH_INEWS_GATEWAY_FILE (Anuj Lal) wrote:
>>>
>>>
>>>
>>> I heard oracle 7.1 has parallel query option.
>>> I have single query which has process table with 10 million rows
>>> and fetch around 50000 rows.
>>> How can I use the paralle query option for this query.
>>> Do i have to declare different cursor for each process fetching data
>>> from same table.
>>> Thanks
>>> alal1_at_eccdb1.pms.ford.com
>>>
>>
>>
>>There are 2 ways that this works. The first is that you alter the
>>table and declair access to it to be parallel. The syntax is
>>specified in the 7.1 documentation ademdum that should have been
>>shipped with your 7.1 version.
>>
>>The second way is to embed a hint in your query something like
>>this:
>>
>>SELECT --+ PARALLEL(table alies, degree of parallelism, number
>> of nodes )
>> * FROM table ...
>>
>>Again, the syntax is sepecifed in the documentation addendum.
>
>Please note that this will only be beneficial if you have multiple processors
>that you can dedicate to this task. A single-CPU machine cannot take advantage
>of parallelism.
>
>Ed
>
Not entirely true. Even on uniprocessor machines, one can take advantage of parallel query. The benefits, ofcourse are not comparable to SMP machines.

-- 
========================================================================
Vipin Gokhale
DEC SBU
Oracle Corporation
Disclaimer:         Any opinions expressed here are mine and not that of
                    my employer.
========================================================================