Hi Marc,

Will it make a difference if you only make text indexing active on predicates 
that are not related to the long DNA sequences? You don't have to enable full 
text indexing on every predicate, it is just the most common way. 

The size of the GSPO indexes may be affected by differences in the size of 
object literals, although I don't know how it is implemented. If you never need 
to do queries that rely on an object literal, you may be able to reduce the 
size of the indexes by removing O based indexes...

Cheers,

Peter





----- Original Message ----
> From: Marc-Alexandre Nolin <[email protected]>
> To: Ceriel Jacobs <[email protected]>
> Cc: [email protected]
> Sent: Fri, 12 March, 2010 7:37:43 AM
> Subject: Re: [Virtuoso-users] compression
> 
> Thanks for the answer, but I would like to keep the indexes but
compress some 
> literals that will never be search of filtered in the
triples store. Those 
> literals will be use if I ask for the triple of
an application X.

For 
> example, DNA sequence can have thousands and thousands of
characters lenght. 
> I want them into the triplestore, but if the were
compress, I would save a 
> lot of space on the 
> hardrive.

Thanks,

Marc-Alexandre

2010/3/11 Ceriel Jacobs 
> <
> href="mailto:[email protected]";>[email protected]>:
> Hi 
> Marc-Alexandre,
>
> Our experience is that the new index(es) that 
> are added to the triple store after upgrading Virtuoso from version 6.0.1 to 
> 6.1.0, increase the database size by a factor of 3.
>
> Our initial 
> was 350MB and became 1GB.
>
> To make the database smaller, and less 
> performant, you could try to remove the extra added indexes by reverse 
> engineering the upgrade instructions: Upgrading to VOS 6.1.0 from an earlier 
> VOS 
> 6.x <
> href="http://virtuoso.openlinksw.com/dataspace/dav/wiki/Main/UpgradingToVOS610#Upgrading%20to%20VOS%206.1.0%20from%20an%20earlier%20VOS%206.x";
>  
> target=_blank 
> >http://virtuoso.openlinksw.com/dataspace/dav/wiki/Main/UpgradingToVOS610#Upgrading%20to%20VOS%206.1.0%20from%20an%20earlier%20VOS%206.x>
> > 
> or to be more precise, have a look in the libsrc/Wi/clrdf23.sql 
> script.
>
>
> Best,
> Ceriel Jacobs
>
> Op 
> 11 mrt 2010, om 21:41 heeft 
> ymailto="mailto:[email protected]"; 
> href="mailto:[email protected]";>[email protected]
>  
> het volgende geschreven:
>
>> Message: 1
>> Date: Wed, 
> 10 Mar 2010 10:09:08 -0500
>> From: Marc-Alexandre Nolin <
> ymailto="mailto:[email protected]"; 
> href="mailto:[email protected]";>[email protected]>
>> Subject: 
> [Virtuoso-users] compression
>> To: 
> ymailto="mailto:[email protected]"; 
> href="mailto:[email protected]";>[email protected]
>> 
> Message-ID:
>>       <
> ymailto="mailto:[email protected]"; 
> href="mailto:[email protected]";>[email protected]>
>> 
> Content-Type: text/plain; charset=ISO-8859-1
>>
>> 
> Hello,
>>
>> I've a N3 dump I'm currently loading into a 
> Virtuoso Server (a
>> complete NCBI Genbank). One literal have always 
> huge size. Its the one
>> related to the predicate "sequence". Is it 
> possible to compress
>> literal with a rule based on 
> predicate?
>>
>> For example:
>> If predicate == 
> sequence then gzip the literals
>>
>> In classical database 
> like MySQL, its possible to compress information
>> in a blob so that 
> the resulting database is smaller. I would like to
>> do the same here. 
> To put this in perspective, the size of the
>> compressed N3 files is 
> 73 GB. I've load 25% of this in the virtuoso.db
>> file already weight 
> around 250 GB. I'm not sure if the full text index
>> is in there with 
> this size.
>>
>> Thanks,
>>
>> 
> Marc-Alexandre 
> Nolin
>
>

------------------------------------------------------------------------------
Download 
> IntelĀ® Parallel Studio Eval
Try the new software tools for yourself. 
> Speed compiling, find bugs
proactively, and fine-tune applications for 
> parallel performance.
See why Intel Parallel Studio got high marks during 
> beta.

> >http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Virtuoso-users 
> mailing list

> href="mailto:[email protected]";>[email protected]

> href="https://lists.sourceforge.net/lists/listinfo/virtuoso-users"; 
> target=_blank 
> >https://lists.sourceforge.net/lists/listinfo/virtuoso-users


 

Reply via email to