Home

Search Engine FAQs

 


How is Sambar Server Search Engine different than AltaVista or Yahoo?

The Sambar Server Search Engine is comparable to AltaVista's search engine, but is able to search and index many fewer pages. The Sambar Server Pro server includes a spider that enables remote sites to be indexed, the standard search engine will only index local files.

How much does Sambar Server Search Engine cost?

The Sambar Server Search Engine is a free, unsupported product, which comes bundled with the Sambar Server. The Sambar Server Pro distribution enables the following functionality: Proximity Search and Search Engine Spider.

What is the difference between 'exact' and 'metaphone' ?

The Sambar Server Search Engine allows documents to be indexed with either an "exact" match algorithm (words must searched must exactly match the indexed words -- unless wildcarding is used), or "metaphone". The metaphone index algorithm provides a soundex-like word match that can accomodate misspellings. The metaphone index algorithm should only be used with english documents and cannot be used with wildcard searches. Furthermore, if the index algorithm is changed, the index must be rebuilt.

Are there programmable APIs for the Sambar Server Search Engine ?

The APIs for the search engine are presently not exposed. Nor can the search engine run outside the Sambar Server.

What file formats does the Sambar Server Search Engine support?

The Sambar Server Search Engine supports ASCII, HTML and Microsoft Word 8 documents (Word 95 and later formats). such as Word or Frame.

Are the Sambar Server Search Engine requests logged ?

Search Engine queries are logged to the file log/search.log if the configuration entry Log Search Requests is set to true.

Does the Sambar Server Search Engine support European languages?

Many users are successfully applying the Sambar Server Search Engine to non-English documents. In general, ISO Latin-1 character sets should yield satisfactory results. There are no plans at the present time to extend the character set beyond single-byte ISO Latin-1.

How do I use the Sambar Server Search Engine in a virtual-host environment ?

At this time, all documents indexed by the Sambar Server Search Engine are assumed to be relative to the Documents Directory. By using separate Search Indexes for each virtual-host, you can use the Sambar Server Search Engine in virtual-host environments. Care must be taken to specify the root directory path of the virtual-host when indexing.

Can the Sambar Server Search Engine index aliased directories, and documents outside the server's document root?

The Sambar Server Search Engine indexes all documents "relative" to the server's document root directory. So while the search engine can be instructed to index the directory c:/foobar, a subsequent search of the index will display the documents relative to the server root. So if the the foobar directory tree appears as:

c:/foobar/index.htm
c:/foobar/special/index.htm
c:/foobar/special/zippy.htm

A search that results yields these files will display the links as:

/index.htm
/special/zippy.htm
/special/index.htm

As you can see, if c:/foobar is not your document root, then requests for /index.htm will result in a Not Found message unless these files/directories are aliased.

IMPORTANT! Because the search engine receives its search directories as a space-separated list of directories, you may not specify a directory that contains a space in it as a directory/path to be searched.

Clarification from Joe Smith...
If the config.ini has "alias /disk-f/ f:/www/disk-f/" then you can tell the Sambar Search Engine to index "f:/www". The resultant index will have:

/disk-f/index.htm
/disk-f/subdir/index.htm

In order for this to work, the documents have to live in a subdirectory that has the same name as the string being aliased, and you tell the indexer to start with the parent of the subdirectory.

In the example above, "f:/www" is a subdirectory that has only one thing in it, the "disk-f" subdirectory, and the name of the subdirectory matches the alias.

Can the Sambar Server Search Engine maintain multiple indexes of documents ?

Yes, you the Sambar Server Search Engine can maintain multiple indexes. In addition, you can search across multiple indexes in a single query by providing a space separated list of the indexes to be searched in the indexname parameter to the query.

Can the Sambar Server Search Engine search multiple web servers?

Yes. The Sambar Pro Server includes the ability to spider web servers to index their content. The freeware search engine is limited to searching documents located on the same physical devices as the Sambar Server.

Can the Sambar Server Search Engine point a search to a specific section of a document?

When the Sambar Server Search Engine returns a list of documents resulting from a query, it returns the entire document. For this reason, the Search Engine is best suited for use with large collections of small (1-2 page) documents, rather than collections of large documents.

Can the Sambar Server Search Engine search on a specific phrase?

The Sambar Server Pro version supports this functionality, known as "proximity" searching. Any phrases placed between quotes must be found within one word of the preceeding word. Note: Only the first 7 occurances of a word in a single file are stored for proximity purposes.

Does the Sambar Server Search Engine index special characters (eg. "&") ?

The list of characters that make up "valid" words is configured in the [search] section of the config.ini file.

Can automatic indexing be set up?

There are no plans to support automatic indexing at this time. This feature is being evaluated for a future release.

What is the maximum number of documents returned from a Sambar Server Search Engine query?

The maximum results from a query is 500 documents.

Can the search engine be used to index RDBMS content ?

Yes. Using the database scripting routines, data can be queried from a relational database and then indexed using the search engine searchadd RPC. By placing the searchadd within the DBMS fetch loop, each row returned can be indexed automatically. (Note: The database script interface limits individual column sizes to 2K.) The following is an example of how a query might be performed:

<RCQtestdb sql="select title, pagelink, content from foobar">
<RCwhile RCFtestdb = 1>
<RCXsearchadd indexname=dbindex title=RCDtestdb.1 url=RCDtestdb.2 data=RCDtestdb.3>
<RCendwhile>

Important: The searchadd RPC may only be executed by the System Administrator. The template file with the above script should be placed in the sysadmin folder to ensure the administrator access.

Can the keywords used in a search be high-lighted in the resulting documents found ?

Yes. The optional highlight parameter can be passed to the find search query that will result in all text/html documents being generated with the /session/highlight argument prepended to the result link. This built-in RPC highlights the keywords passed to it in the color specified. The following illustrates how to use this feature:

<RCXfind query="Sambar Server" highlight=red>

Is there a way that the search can be set for the filename versus content?

If you have a catalog of images (stored in an RDBMS) along with titles and descriptions of the images that you would like to index to allow for searching of the images, you build an index of these images by using the searchadd RPC. Assuming you have an RDBMS table with the following contents:

titlefilenamedescription
Image 1 Title/image1.gifImage 1 description...
Image 2 Title/image2.gifImage 2 description...
Image 3 Title/image3.gifImage 3 description...
Image 4 Title/image4.gifImage 4 description...

Create a script (.stm) page in the /sysadmin folder (searchadd must be run as the system administrator) that queries the table and generates the searchadd entries for each row (note, you can schedule this scripted page to regenerate the index periodically!):

Free all entries for the Image search index...

<RCXsearchfree indexname=imagesearch>

Query for each row and create the search entry... 

<RCQmydb sql="select title, filename, description from myimages">

<RCwhile RCFmydb=1>
	<RCXsearchadd indexname=imagesearch title=RCDmydb.1 url=RCDmydb.2 data=RCDmydb.3>
<RCendwhile>

© 1998-2000 Sambar Technologies. All rights reserved. Terms of Use.