Full-Text Searching

Thread view:

Enable EMail Alerts

Start New Thread

Thread rating:

Cindy - 26 Feb 2008 18:55 GMT

Hi everyone,

I was wondering whether anyone had any experience designing Full-Text
Search web sites. More specifically, I am wondering what the
advantages are of using either SQL Server's Full-Text or a third party
application, such as Lucene.net ( http://www.dotlucene.net/ ).

I would like to search:

* Word Files (Lots)
* A field or two in a large table (three hundred thousand records)
* So on and so forth

Any considerations for gaining speed, or perhaps the suggestion of
another database (PHP with MySQL or whatever) would be helpful.

Thank you and regards,

Cindy

Reply to this Message

Cindy - 27 Feb 2008 04:29 GMT

> Hi everyone,
>
[quoted text clipped - 15 lines]
>
> Cindy

Anyone there? Are there any 'MVP's or experts there who can answer?

Reply to this Message

Ken Fine - 28 Feb 2008 04:38 GMT

Hi there Cindy,

I run a large-ish news site that uses Index Server, which is Ancient History
but works quite well for my applications.

Index Server is free and will index many different content types: PDFs,
HTML, etc, unlike SQL Server Full Text Search, which I previously used. For
indexing dynamically generated web page content I have systems that render
out our dynamic pages as static webpsages which are subsequently indexed. I
then have a search results page I programmed that directs people to the live
content based on what they found in the static indexed content.

Index Server has a lot of features but the documentation is, uh,
"distributed." Read widely.

Why Microsoft has not maintained/updated/documented IS more effectively is a
mystery to me. I'm at the threshold of looking at third party solutions or
maybe building something using Google's search APIs. This is very silly:
with all of the other pretty awesome server tech that Microsoft makes, it's
a shame that there search solutions aren't so well maintained.

Somebody correct me if I'm wrong here. My understanding is that Sharepoint
Server had an industrial-strength indexer that shipped with it and I think I
remember some consideration being given to this becoming a standalone
product. Is it?

-KF

> Hi everyone,
>
[quoted text clipped - 15 lines]
>
> Cindy

Anyone there? Are there any 'MVP's or experts there who can answer?

Reply to this Message

Ken Fine - 28 Feb 2008 04:57 GMT

If you're looking at ASP.NET and IShave a look at this:
http://www.codeproject.com/KB/aspnet/search.aspx

> Hi everyone,
>
[quoted text clipped - 15 lines]
>
> Cindy

Anyone there? Are there any 'MVP's or experts there who can answer?

Reply to this Message

Alexey Smirnov - 28 Feb 2008 08:32 GMT

> Hi everyone,
>
[quoted text clipped - 8 lines]
> * A field or two in a large table (three hundred thousand records)
> * So on and so forth

Index Server and SQL Server can do all this and much more.
http://www.google.com/search?q=index+server+sql+server+asp.net&hl=en

> Any considerations for gaining speed, or perhaps the suggestion of
> another database (PHP with MySQL or whatever) would be helpful.

As Ken already mentioned there are two new products from Microsoft

Microsoft Search Server 2008
Microsoft Search Server 2008 Express (free)

http://www.microsoft.com/enterprisesearch/serverproducts/searchserver/
http://www.microsoft.com/enterprisesearch/serverproducts/searchserverexpress/

I would not recommend them for the internet sites, but you can try

Reply to this Message

Cindy - 29 Feb 2008 03:24 GMT

> > Hi everyone,
>
[quoted text clipped - 22 lines]
>
> I would not recommend them for the internet sites, but you can try

Hello and thanks for the previous replies. Am I to understand that
SQL Server Full-Text Search can not index HTML, PDF, or Word
documents? What I need is a relational structure, meaning a user
(from the USER table, presumably) will have several documents
associated to him. Once I find the searched words in a file I get
redirected to the associated user. Can't I just store each file in an
image field in SQL Server and query it from there? In other words,
Index Server would be fine if I only had to search my web site or a
particular directory, but can it handle "related" (One user-->Many
files) information?

Thanks for all your help.

Cindy

Reply to this Message

Alexey Smirnov - 29 Feb 2008 07:49 GMT

> Hello and thanks for the previous replies. Am I to understand that
> SQL Server Full-Text Search can not index HTML, PDF, or Word
[quoted text clipped - 3 lines]
> redirected to the associated user. Can't I just store each file in an
> image field in SQL Server and query it from there?

Hello Cindy,

yes, you can use a full-text search if your files are stored in the
image field. In this case you don't need any other software, except
the iFilter for PDF indexing. Adobe PDF iFilter lets you index PDF
documents with SQL Server. Starting with Acrobat Reader 7, iFilter
functionality is bundled within the Acrobat Reader and you must
install it (current version is 8) on the server where SQL Server is
installed.

http://www.google.com/search?hl=en&q=full+text+search+files+image+field+sql+server

> Index Server would be fine if I only had to search my web site or a
> particular directory,

Yes, exactly.

Hope this helps

Reply to this Message

Cindy - 01 Mar 2008 05:00 GMT

> > Hello and thanks for the previous replies. Am I to understand that
> > SQL Server Full-Text Search can not index HTML, PDF, or Word
[quoted text clipped - 22 lines]
>
> Hope this helps

Say, that's all been very helpful. But now it's time to get to work!
There are a myriad of things to learn; however, once I get something
scalable going I can build from there. It is all really just a set a
very simple problems.

Thanks again!

Reply to this Message

Rate this thread:

Full-Text Searching

Free Magazines