Hi David,
due to the complexity of XQuery, it’s often a big challenge for the
query optimizer to find out if an index can be used or not. This is
the reason why some XML database don’t even try to rewrite queries for
index access.
What you can always do is: directly address the available index
structures via the Full-Text and Index modules [1,2]. This will give
you the guarantee that the index is indeed utilized.
I'm sure there are also ways to rewrite your existing query in order
to ensure index access, but I’d first need some more time to get into
this. As a general note, you can help the optimizer by directly
specifying the addressed database in your query. The following two
examples may illustrate this. Query A und B will be optimized for
index access, while Query C won’t (well, at least not in BaseX ≤ 7.5):
Query A
db:open("db")//*[text() contains text { "A","B" }]
Query B
let $terms := ("A", "B")
for $x in db:open("db")//*
where $x/text() contains text { $terms }
return $x
Query C
declare function local:x($db, $terms) {
db:open($db)//*[text() contains text { $terms }]
};
let $terms := ("A", "B")
let $db := "db"
return local:x($db, $terms)
Best,
Christian
[1] http://docs.basex.org/wiki/Full-Text_Module
[2] http://docs.basex.org/wiki/Index_Module
___________________________
On Wed, Jan 2, 2013 at 10:59 PM, David Stuebe <dstu...@asascience.com> wrote:
>
> Hi Chris
>
> I did look at the Query Info in the Gui and it seems you are correct. The
> complex query I created is not utilizing the full text index.
>
> Details here:
> https://gist.github.com/4438473
>
> The gist contains query info for the complex search that I created and for a
> simple one that just executes the 'contains text' clause.
>
> The compile statement for the simple query includes the following line:
> "applying full-text index"
> The complex query does not…
>
>
> So what to do now? How do I figure out why basex is not using the full text
> index as the initial step in limiting the results of my search?
>
> Thanks
>
> David
>
>
> From: "basex-talk-requ...@mailman.uni-konstanz.de"
> <basex-talk-requ...@mailman.uni-konstanz.de>
> Reply-To: "basex-talk@mailman.uni-konstanz.de"
> <basex-talk@mailman.uni-konstanz.de>
> Date: Tue, 1 Jan 2013 06:00:02 -0500
> To: "basex-talk@mailman.uni-konstanz.de"
> <basex-talk@mailman.uni-konstanz.de>
> Subject: BaseX-Talk Digest, Vol 37, Issue 1
>
> Message: 1
> Date: Mon, 31 Dec 2012 20:56:01 +0100
> From: Christian Gr?n <christian.gr...@gmail.com>
> To: David Stuebe <dstu...@asascience.com>
> Cc: Kyle Wilcox <kwil...@asascience.com>,
> "basex-talk@mailman.uni-konstanz.de"
> <basex-talk@mailman.uni-konstanz.de>
> Subject: Re: [basex-talk] Initial xquery for full text and geospatial
> search of ISO 19115 docs
> Message-ID:
> <cap94bnpneuwgumgl-k2zmw6gcd4wqnc82kbr5es8h+brz1f...@mail.gmail.com>
> Content-Type: text/plain; charset=windows-1252
>
> Hi David,
>
> thanks for the insight into your project. It may be that the full-text
> index is not utilized by your XQuery expression. Did you have a look
> at the query info (e.g. via GUI, InfoView, or -V on command line)? Did
> you manage to write simpler queries that are processed faster?
>
> Best,
> Christian
> ___________________________
>
> On Mon, Dec 31, 2012 at 7:31 PM, David Stuebe <dstu...@asascience.com>
> wrote:
>
>
>
>
> Hi Basex Folks
>
> I have written a simple minded xquery script which can be used in a post to
> search ISO 19115 metadata documents.
> As I am a newbie to xquery and basex I expect that is much that I could do
> to improve performance, but currently searches take up to 15 seconds. The
> server hardware is not blazing fast and we are running inside tomcat? but
> that is certainly not acceptable for a database which is still quiet small.
> Initially it was fine, responding in less than a second, but when we
> expanded from less than 100 docs to about 35000 the time grew at least
> linearly. I am hoping this is due to my poor xquery programming or a setting
> on the server.
>
> Here is the query we are running using a post request with the declared
> variables at the bottom filled in by the UI:
> https://github.com/asascience-open/glos_catalog/blob/master/queries/full_search.xml
>
> Here are some of the ISO xml documents that we have in our basex DB:
> https://github.com/asascience-open/glos_catalog/tree/master/ISOs
>
> It is used by this site: http://explorer.glos.us/
> To provide geospatial metadata search. Fill in a text value like "water" or
> "temperature" in the search box in the top right?
>
> The server is here: http://64.9.200.113:8080/BaseX73/
> User Name: user
> Password: glos
> ACL: read only
>
>
> I am hoping it is as simple as a setting on the basex server. Here is the
> current server info:
>
> info database
> Database Properties
> Name: glos
> Size: 422 MB
> Nodes: 17045388
> Documents: 34922
> Binaries: 0
> Timestamp: 27.12.2012 00:58:03
>
> Resource Properties
> Timestamp: 22.12.2012 01:43:05
> Encoding: UTF-8
> Whitespace Chopping: ON
>
> Indexes
> Up-to-date: true
> Text Index: ON
> Attribute Index: ON
> Full-Text Index: ON
>
> The index info is available here:
> https://github.com/asascience-open/glos_catalog/blob/master/queries/glos_index_info.txt
>
>
>
> David Stuebe
>
> Scientist & Software Engineer ? RPS ASA
>
> 55 Village Square Drive
> South Kingstown, RI 02879-8248
>
> Tel: +1 (401) 789-6224
>
> Email: david.stu...@rpsgroup.com
> www: asascience.com | rpsgroup.com
>
> A member of the RPS Group plc
>
>
> _______________________________________________
> BaseX-Talk mailing list
> BaseX-Talk@mailman.uni-konstanz.de
> https://mailman.uni-konstanz.de/mailman/listinfo/basex-talk
>
>
>
>
> _______________________________________________
> BaseX-Talk mailing list
> BaseX-Talk@mailman.uni-konstanz.de
> https://mailman.uni-konstanz.de/mailman/listinfo/basex-talk
>
_______________________________________________
BaseX-Talk mailing list
BaseX-Talk@mailman.uni-konstanz.de
https://mailman.uni-konstanz.de/mailman/listinfo/basex-talk
