Compiling a Thesaurus File - Full-Text Retrieval (FTR) - Help

Full-Text Retrieval (FTR) Help

Language
English
Product
Full-Text Retrieval (FTR)
Search by Category
Help

The fthmake utility compiles the thesaurus by reading the thesaurus source file and creating a thesaurus object. If you are using a character variant rules file in addition to the thesaurus file, you must ensure that the thesaurus lookup includes any typographical rules duplicated in the rules file. Use this utility to incorporate any duplicated rules into the thesaurus object file. The software will ignore the duplicates in the variant rules file.

Start fthmake with this syntax:

fthmake sourcefile objectfile [-f filter_list][-l rulesfile]

where:

  • sourcefile is the name of the thesaurus source file to be compiled (it must contain the three-character extension .FTS).

  • objectfile is the name of the object file to be created (it must have a filename containing eight or fewer characters plus the three- character extension .FTH).

  • filter_list specifies a filter list to be applied to the thesaurus source file.

The fthmake utility does not support any redirection of input or output. Use an explicit value for the sourcefile or objectfile on the command line.

The rulesfile is the name of the file that contains the variant rules you want to incorporate into the thesaurus object file. When this optional parameter is specified, each word found on the LHS of a thesaurus source rule is subject to character variant expansion according to the variant rules in the rulesfile. This enables typographic variants of search terms to be eligible for thesaurus expansion.

For example, to build a thesaurus file called USER.FTH, enter:

fthmake user.fts user.fth

The default filter list (nti:s) can be overridden by the -f parameter. If the translation filter is selected, it translates the source text from the external character set to the OpenText Internal Character Set (FTICS) equivalent. If the source text is already represented in the FTICS, a specification for the standard filter (s) should be used.

If there are any compilation errors, fthmake writes a message to standard error before it exits. For example, occasionally a source file can't be read, or an object file cannot be created due to access protection. A syntax error in the source file, such as a missing semicolon character, would write a message (for example, words at end of file or need right terminator). If a suffix rule is entered in the source file more than once, the message suffix root is in twice is produced.

Any problem encountered in writing to any part of object file (including running out of space in the file system) produces the message can't write object file. If an object file has been created, but writing was not completed due to an error, it is removed.

For more information about character variant rules files, see Creating a Character Variant Rules File.