Core Files

The base name of the core files is formed by appending a period and 3-character name extension to the 1- to 8-character database name specified at creation time. Core files are binary and accessible only via DtSearch programs.

The DtSearch core files are as follows:

dbname.dbe

Database dictionary file. Binary schema created by dtsrcreate from dtsearch.dbe. Never modified thereafter.

dbname.k00

Main key file for database documents. Created and initialized by dtsrcreate, updated by dtsrload. Contains the b-tree of unique keys for each document.

dbname.k01

Optional key file for database documents. Created and initialized by dtsrcreate. Contains the b-tree of optional keys for each document. Not currently used.

dbname.d00

Documents header file. Created by dtsrcreate, updated by dtsrload. Contains the databases configuration status record and, for each document in the database, a header record and one or more abstract records.

dbname.d01

Compressed text file. Created by dtsrcreate, but updated by dtsrload only for AusText type dataases. Repository of compressed text for each document.

dbname.k21,

dbname.k22, dbname.k23" 10 Key files for words and stems. Created and initialized by dtsrcreate, updated by dtsrindex. Contains the b-tree of each word and stem indexed for the database. The k21 file finds "short" words, 1 to 15 bytes, in the d21 file. The k22 file finds "long" words, 16 to 39 bytes, in the d22 file. The k23 file finds "huge" words, 40 to 133 bytes, in the d23 file. Long and huge word files may not be used depending on the database maximum word size specified at creation time.

dbname.d21,

dbname.d22, dbname.d23" 10 Data files for words and stems. Created and initialized by dtsrcreate, updated by dtsrindex. For each word contains document counts, offset to inverted index (d99 file), and storage recovery data. The d21 file contains short words, the d22 file contains long words, and the d23 file contains huge words. Long and huge word files may not be used depending on the database maximum word size specified at creation time.

Language Files

Databases also need a set of files associated with the DtSearch language of the database. When looking for these files DtSearch will first look for a customized version applicable only to a database, and then look for the generic language version. Like core files, the base file name of a customized language file is formed by the database name and a 3 character extension. The alternative generic language files are named with a language name and the same 3 character extension.

Language files are mandatory or optional depending on the language. See dtsrlangfiles(4) for formats of language files.

The DtSearch language-related files are as follows:

dbname.stp

Stop file. The supported stop files are:

Stop lists are mandatory for European languages, and optional for other supported languages.

dbname.inc

An include list is always optional for all supported languages. There are no generic versions of include lists.

eng.sfx

ForDtSrLaENG and DtSrLaENG2. and is not currently required for other supported languages.

dbname.knj

jpn.knj for DtSrLaJPN2. A kanji compounds file is mandatory only for language number 7 DtSrLaJPN2, a supported Japanese language.

Examples

Files associated with a minimum DtSrLaENG database (English, ASCII) that uses no customized or optional files:

All core files plus eng.stp, eng.sfx.

Files for a DtSrLaITA database (Italian, ISO Latin-1) with enhanced stop list and an include list:

All core files plus dbname.stp, dbname.inc.

Files associated with a minimum DtSrLaJPN database (Japanese with full, automatic kanji compounding) that uses no customized or optional files:

Only core files.

Files for a DtSrLaJPN2 database (Japanese with kanji compounds from a word list), with optional stop list for ASCII substrings:

All core files plus dbname.stp, jpn.knj.