You are viewing a plain text version of this content. The canonical link for it is here.
Posted to xindice-users@xml.apache.org by asif ali <as...@yahoo.com> on 2002/11/25 20:24:32 UTC

indexing

hi

I was just experimenting with XIndice and stored a
complete set of documents (abt 4 MB)
and tried to get all the elements/nodes

I applied indexing also but even them it is very slow

I have abt 18000 nodes of various types
and it took 3 minutes to qry all

where as my Parser application takes abt 25 seconds ..

Is there any way to improve indexing ..

asif
--- Dave Viner <dv...@yahoo-inc.com> wrote:
> try
> 	xindice ld -c /db/foo/bar
> 
> xindice -h should show you the usage information for
> the command line
> client.
> 
> dave
> 
> 
> -----Original Message-----
> From: asif ali [mailto:asifjaan@yahoo.com]
> Sent: Monday, November 25, 2002 3:57 AM
> To: xindice-users@xml.apache.org
> Subject: how to list doc in a collection
> 
> 
> how can we list doc in collection from command line
> thanks
> 
> 
> __________________________________________________
> Do you Yahoo!?
> Yahoo! Mail Plus  Powerful. Affordable. Sign up
> now.
> http://mailplus.yahoo.com
> 
> 


__________________________________________________
Do you Yahoo!?
Yahoo! Mail Plus � Powerful. Affordable. Sign up now.
http://mailplus.yahoo.com

Re: indexing

Posted by asif ali <as...@yahoo.com>.
hi
Thanks
well I tried out indexing based on element nodes

as I was running queries to get the individual nodes
and make objects 
//LogicalPart
//Box

and I want indexing to be done based on these nodes...

so I want to index all LogicalPart s and say all Box s

Using Xerces Parser for C++ takes 20 seconds to read
all the elements and make objects from the data

and I tried Xindice today and simple reading out all
nodes (by executing queries takes 3 min)

I know in parser I am just reading the doc and here I
am excuting at least 10-12 queries invloving all docs
(27 in total)

i dont want it to be very fast bcz it is not
requirement .. i intend exploiting other features
which are not there in Parsing application ...

I would love to use dunamic linking so that I can
query the query results (although not sure how it can
be done)

but I would like it to be little fast then 3 minutes
on P4 machine


asif

--- Jeff Greif <jgrehi
if@alumni.princeton.edu> wrote:
> It helps to be more specific in your question.
> 
> Did you use the query string "//*" on your
> collection?
> 
> In general, indexing will be completely useless with
> this query.  Indexing
> maps element and attribute names or values to
> documents in the collection.
> If there are no restrictions in your query, the
> presence of an index either
> has no effect, or negative effect, depending upon
> the details of
> implementation.  In the current xindice, the index
> is used to pick out
> subsets of the documents that are known to contain
> matches to the query, but
> then an XPath processor must still run on each
> selected document.  Use of
> the descendant axis may involve a lot of computation
> in this phase.
> 
> An index is good for queries like
>   "/a/b/c"  (all the c elements which are children
> of /a/b)
> or
>  "/a/b/c[@code='7']"
> 
> Jeff
> ----- Original Message -----
> From: "asif ali" <as...@yahoo.com>
> To: <xi...@xml.apache.org>
> Sent: Monday, November 25, 2002 11:24 AM
> Subject: indexing
> 
> 
> > hi
> >
> > I was just experimenting with XIndice and stored a
> > complete set of documents (abt 4 MB)
> > and tried to get all the elements/nodes
> >
> > I applied indexing also but even them it is very
> slow
> >
> > I have abt 18000 nodes of various types
> > and it took 3 minutes to qry all
> >
> > where as my Parser application takes abt 25
> seconds ..
> >
> > Is there any way to improve indexing ..
> >
> > asif
> > --- Dave Viner <dv...@yahoo-inc.com> wrote:
> > > try
> > > xindice ld -c /db/foo/bar
> > >
> > > xindice -h should show you the usage information
> for
> > > the command line
> > > client.
> > >
> > > dave
> > >
> > >
> > > -----Original Message-----
> > > From: asif ali [mailto:asifjaan@yahoo.com]
> > > Sent: Monday, November 25, 2002 3:57 AM
> > > To: xindice-users@xml.apache.org
> > > Subject: how to list doc in a collection
> > >
> > >
> > > how can we list doc in collection from command
> line
> > > thanks
> > >
> > >
> > >
> __________________________________________________
> > > Do you Yahoo!?
> > > Yahoo! Mail Plus  Powerful. Affordable. Sign up
> > > now.
> > > http://mailplus.yahoo.com
> > >
> > >
> >
> >
> > __________________________________________________
> > Do you Yahoo!?
> > Yahoo! Mail Plus - Powerful. Affordable. Sign up
> now.
> > http://mailplus.yahoo.com
> >
> 


__________________________________________________
Do you Yahoo!?
Yahoo! Mail Plus � Powerful. Affordable. Sign up now.
http://mailplus.yahoo.com

Re: indexing

Posted by Jeff Greif <jg...@alumni.princeton.edu>.
It helps to be more specific in your question.

Did you use the query string "//*" on your collection?

In general, indexing will be completely useless with this query.  Indexing
maps element and attribute names or values to documents in the collection.
If there are no restrictions in your query, the presence of an index either
has no effect, or negative effect, depending upon the details of
implementation.  In the current xindice, the index is used to pick out
subsets of the documents that are known to contain matches to the query, but
then an XPath processor must still run on each selected document.  Use of
the descendant axis may involve a lot of computation in this phase.

An index is good for queries like
  "/a/b/c"  (all the c elements which are children of /a/b)
or
 "/a/b/c[@code='7']"

Jeff
----- Original Message -----
From: "asif ali" <as...@yahoo.com>
To: <xi...@xml.apache.org>
Sent: Monday, November 25, 2002 11:24 AM
Subject: indexing


> hi
>
> I was just experimenting with XIndice and stored a
> complete set of documents (abt 4 MB)
> and tried to get all the elements/nodes
>
> I applied indexing also but even them it is very slow
>
> I have abt 18000 nodes of various types
> and it took 3 minutes to qry all
>
> where as my Parser application takes abt 25 seconds ..
>
> Is there any way to improve indexing ..
>
> asif
> --- Dave Viner <dv...@yahoo-inc.com> wrote:
> > try
> > xindice ld -c /db/foo/bar
> >
> > xindice -h should show you the usage information for
> > the command line
> > client.
> >
> > dave
> >
> >
> > -----Original Message-----
> > From: asif ali [mailto:asifjaan@yahoo.com]
> > Sent: Monday, November 25, 2002 3:57 AM
> > To: xindice-users@xml.apache.org
> > Subject: how to list doc in a collection
> >
> >
> > how can we list doc in collection from command line
> > thanks
> >
> >
> > __________________________________________________
> > Do you Yahoo!?
> > Yahoo! Mail Plus  Powerful. Affordable. Sign up
> > now.
> > http://mailplus.yahoo.com
> >
> >
>
>
> __________________________________________________
> Do you Yahoo!?
> Yahoo! Mail Plus - Powerful. Affordable. Sign up now.
> http://mailplus.yahoo.com
>