 |
Full-text and Domain Search variables
Updated: 05/01/2002
Related link:
More Ask Professor INI |
 |
This month, the Professor reviews a host of NOTES.INI settings for full-text indexing, Domain Search, and the Domain Index (the central index of all the text in all the databases and all the file systems selected for inclusion in Domain Search). Most of these settings are not covered in the core documentation shipped with Domino Release 5.0.3, as they were introduced in subsequent Maintenance Releases (MRs).
Full-text variables
First, let's look at the full-text variables.
FT_Alternate_Filter
In release 5.0.3, an alternate filter is provided for attachment indexing that enhances the capabilities of the Keyview filter. The alternate filter provides support for file types, such as Ichitaro file format, that are important internationally. You enable this alternate filter by using the NOTES.INI parameter FT_Alternate_Filter=naltfltr.dll.
FT_Index_Attachments
You can employ the NOTES.INI variable FT_Index_Attachments=2 to turn off attachment indexing. This overrides any attachment indexing option set up directly in a Notes database or the domain indexer. (FT_Index_Attachments=1 is an unconditional command to index all attachments.)
FT_Index_Ignore_Attachment_Types
You can use FT_Index_Ignore_Attachment_Types to tell the full-text indexer to ignore specified attachment types. Since release 5.0.5, Notes (by design) ignores the following types of files when indexing attachments: NSF, NTF, SYS, PAG, IMG, WAV, GIF, JPG, DLL, TAR, EXE, ZIP, AU, MPG, MOV, and MP3. The NOTES.INI setting FT_Index_Ignore_Attachment_Types debuted in 5.0.5, and it lets you specify additional file types to ignore in the indexing of attachments. For example, the line FT_Index_Ignore_Attachment_Types=*.tif tells the indexer not to index attachments with the TIF extension. This can reduce the size of the index and hasten the return of search results. If you want to specify multiple file types to be ignored, separate them with a comma.
FT_Index_Accent_Sens
You can employ the setting FT_Index_Accent_Sens=1 to make the full-text index sensitive to accented characters. This helps, for example, the user who connects to a Domino Web server with Internet Explorer and wants to search a Notes database for Nordic characters, such as Æ, Å, Ø.
FTG_Index_Limit
Domino's default maximum size for a full-text index is 6 MB (1 MB for the Mac). You can increase this maximum size using the NOTES.INI setting FTG_Index_Limit=n, where n is the maximum size for a full-text index (in bytes). For example, to increase the maximum size for a full-text index to 12 MB, add the line FTG_Index_Limit=12582912 to the NOTES.INI file and restart the server. (12 MB is calculated as 12*1024*1024 = 12582912.)
Update_No_Fulltext
Use the setting Update_No_Fulltext=1 to turn off full-text indexing on a server. Set this variable to 0 to turn full-text indexing on.
FT_LIBNAME
The Global Text Retrieval (GTR) search engine GTR-34 has been part of the Domino server installation since 5.0.3, but it was not enabled by default. To turn it on in releases 5.0.3 and 5.0.4 (and for 5.0.5/5.0.6 servers on the AS/400 platform), add the following line to the server's NOTES.INI file:
FT_LIBNAME=ftgtr34
You must then restart the Domino server. Note that if you activate the "alternate" search engine using this NOTES.INI parameter, Domino rebuilds the indices for all databases that are currently indexed. If the parameter is later removed from the NOTES.INI file, all indices are rebuilt in the original R5 format. (Remember that rebuilding indices for large databases can take time and plan accordingly.)
Beginning with Domino 5.0.5, the GTR-34 engine became the default search engine in Domino, so there is no need to use this NOTES.INI setting in that version and later to specify the GTR engine. However, to revert to using the older engine in Domino 5.0.5 or later, you can add the following line to the NOTES.INI file on any client or server:
FT_LIBNAME=ftgtr
FTG_No_Summary
The Inxight Summarizer is a feature for Domain Search on Windows NT that automatically pulls the best n sentences out of a document and displays them at search time as the document summary. This feature is on by default with the GTR-34 search engine. To disable the Summarizer, use the setting FTG_No_Summary=1.
FT_Summ_Default_Language
If this setting does not appear in the NOTES.INI file, the native language from the locale is used as the default language for the Summarizer. If it is set to the NULL string, English is used as the default language. Additionally, it can be set to any language, such as Dutch. Whenever the Summarizer assigns a language to a document, and the Summarizer is unable to summarize in that language, the Summarizer attempts to summarize using the default language instead.
Please note that if the native language is one that is not yet supported, such as Japanese, FT_Summ_Default_Language should be set to NULL or English.
Whenever the Summarizer cannot summarize a document because it is too short or because it cannot summarize in the language specified, neither an error message nor a summary is produced.
FT_Max_Search_Results
Use FT_Max_Search_Results to restrict the number of full-text hits.
By default, the full-text search engine has a limit of 5,000 documents. The functionality of how that 5,000 document limit is applied differs between the new GTR-34 search engine and the prior version of the search engine. With the prior version of the search engine, the 5,000 document search limit did not limit the number of documents returned by the agent search builder; if 100,000 documents matched, the agent acted on all 100,000 documents. However, the GTR-34 search engine does enforce the 5,000 document limit in agent searches. Regardless of how many documents match the search criteria, only 5,000 documents are acted upon by the agent.
Domain Search settings
Domain Search expands the concept of full-text search beyond the individual Domino database and lets you use the Notes client or a Web browser to search every server in a domain for information, whether the information is contained in a Domino database or in another file system. You don't need to know specifically which database or document contains the information you are looking for when Domain Search is set up.
FT_Domain_Directory_Name
By default, the Domain Index is located in the Domino data directory of the Domain server and has the name FTDOMAIN.DI. Use the setting FT_Domain_Directory_Name=path in the server's NOTES.INI file to set the location and name of the Domain Index.
FT_HTML_Title
When it indexes HTML files for Domain Search, Domino indexes only the text contained in the <BODY> tag. To include the text in all HTML tags (for example the <META> or <TITLE> tags), you must first edit the FORMATS.INI file in the Notes or Domino directory by remarking out the 210=htm statement:
REM 210=htm
Then add the following line to the NOTES.INI file:
FT_HTML_Title=1
FT_DOMAIN_IDXTHDS
Use this setting to specify the number of indexing threads used for Domain Search. Use more threads to let the Domain Catalog server index more files simultaneously, but note that this requires more CPU utilization and response to search queries may be slow. With fewer indexing threads, search speeds up because of greater CPU availability, but changes are not reflected in the index as quickly.
Without this setting, the default number of threads used is two threads per CPU. For example, a server with two CPUs uses four indexing threads by default when indexing. You should not exceed eight threads per server or you may degrade the performance of the server, even on servers with more than four CPUs.
Do you have a NOTES.INI questions? Send it in to Professor INI. We'll answer as many questions as we can in future "Ask Professor INI" columns. Keep in mind, however, that we may not be able to answer every question, nor can we quickly respond to requests that require immediate attention. If you need an immediate response to a question, we recommend you post it in the Notes/Domino Gold Release Forum where someone from the general Notes community might be able to help, or contact Lotus Support Services.
|
|