Oracle FAQ | Your Portal to the Oracle Knowledge Grid |
Home -> Community -> Usenet -> c.d.o.tools -> Re: Compressed Indexes fail
I have checked metalink for
compress index 12421
There is a bug 1468034, fixed in 9.0.0
relating to compressed indexes and
rebuilds but it seems to be more about
AND-EQUAL paths with compressed indexes;
but there is a note about a patch on Sun
There is also a bug 1484218 in a similare vein on HP-UX, with Oracle 8.1.6.2
-- Jonathan Lewis Yet another Oracle-related web site: http://www.jlcomp.demon.co.uk Practical Oracle 8i: Building Efficient Databases Publishers: Addison-Wesley More reviews at: http://www.jlcomp.demon.co.uk/book_rev.html Mark Jenkins wrote in message ...Received on Tue Jan 30 2001 - 02:35:11 CST
>It is an index-only query.
>
>I will have to try using a combination of non-indexed and indexed fields
>in the where clause and see what happens. I'm betting the qury will work
>as we can perform the query using the indexes individually and we get
>rows returned.
>
>Thanks.
>
>Mark Jenkins
>
>
>In article <980802024.4244.0.nnrp-14.9e984b29_at_news.demon.co.uk>,
>"Jonathan Lewis" <jonathan_at_jlcomp.demon.co.uk> wrote:
>
>> I have recently seen mention of a bug with
>> compressed indexes in Metalink - can't remember
>> the details but it might have been to do with
>> queries that are index-only queries which this
>> presumably is.
>>
>>
>> --
>> Jonathan Lewis
>> Yet another Oracle-related web site: http://www.jlcomp.demon.co.uk
>>
>> Practical Oracle 8i: Building Efficient Databases
>>
>> Publishers: Addison-Wesley
>> More reviews at: http://www.jlcomp.demon.co.uk/book_rev.html
>>
>>
>>
>> Mark Jenkins wrote in message ...
>> >If we use the default compressed indexes, we get the following error:
>> >
>> >ORA-00600: internal error code, arguments: [12421], [], [], [], [], [],
>> >[],
>> >[]
>> >Current SQL statement for this session:
>> >select count(*) from f_event where eventDate ='22-jan-2001' and
>> >eventtype_key='1'
>> >
>> >If we alter the indexes to NOCOMPRESS everything works fine.
>> >
>> >Is this a known problem/bug?
>> >
>>
>>
>>