]]> dtsrcreateuser cmd dtsrcreateCreate and initialize a DtSearch database dtsrcreate−q−o−fd−fa−aabstr−ddir−wnmin−wx max−llang dbname The dtsrcreate command creates and initializes an instance of a DtSearch database. A DtSearch database consists of a set of related files. If the specified database already exists, after prompting for confirmation, dtsrcreate will erase and reinitialize the preexisting database. The dbname argument is the database name. It is a 1 to 8 ascii character string used at creation time as a base file name, and as a general database identifier thereafter. All created database files are named by assembling the base name, plus a period and a 1 to 3 ASCII character suffix. The database names dtsearch and austext are reserved and may not be specified. The dbname argument can include an optional path prefix. If it does, the database files will be created and initialized in the specified target directory. If no path prefix is specified, the target directory is the current working directory. One of the created database files is based on a model file, dtsearch.dbe, provided with DtSearch. Database creation will fail if the model file cannot be found. dtsrcreate looks for the model file first in the directory specified by a command line option, if any; secondly in the current working directory; and thirdly in the optional dbname target directory. DtSearch databases can be customized with a number of configuration options that are specified only at creation time. Initialization consists of loading into the database a configuration and status record identifying the configuration options for the particular database instance. After initialization, dtsrcreate prints a small report of the current contents of the configuration record to stdout. (See also &cdeman.dtsrdbrec;, which prints the report without changing the database). The customizable features available at database creation time fall into clusters of related capabilities that constitute a set of basic database types. When you select a database type, you prespecify a number of features that are optimized for the basic type of database you want. In the DtSearch database type, documents are not stored in a repository and are not available from the search engine after a search. The abstract returned from a search typically contains a document reference, usually the file name, and the application is itself responsible for accessing the document. Hilighting of search words is possible when the application passes the document cleartext back to the DtSearch API. In an AusText database type, compressed documents are stored directly into a repository and the originals are thereafter ignored. The abstracts returned from searches are typically descriptive of the documents they represent, and are displayed directly to users. Documents can be retrieved from an AusText type database through the API, and the search words are highlighted as desired. The following options are available: If an option takes a value, the value must be directly appended to the option name without white space. −q Suppresses printing of configuration record report. −o Suppresses overwrite prompt; preauthorizes erasure and reinitialization of preexisting database. −ddir Specifies where to find the model dtsearch.dbe file, rather than in the current working directory or target directory. −fd Configure a DtSearch type database. This is the default. −fa Configure an AusText type database. −aabstr Set the maximum abstract size to abstr bytes. This is the maximum permitted length in characters for an abstract string. To optimize space considerations in the database the choice for abstract length may be adjusted upward. Default size depends on the specified database type. (See &cdeman.dtsrfzkfiles; and &cdeman.DtSearch; for more information about abstract fields.) −wnmin Change minimum word size to min characters. This is the minimum word size in characters to be indexed in the database. Document and query words shorter than the minimum are treated as stop list words (see &cdeman.dtsrfzkfiles;). The minimum can be overridden for specific individual words by adding them to the optional include list file (see &cdeman.dtsrfzkfiles;). For most natural languages the default minimum word size is usually correct; permitting very short words will usually cause a significant increase in the storage requirements for the database. This option is typically applicable to single-byte European languages and may be ignored by multibyte language processors. (See &cdeman.DtSearch; for more information about DtSearch word sizes). −wxmax Change maximum word size to max characters. This is the maximum word size in characters. Smaller is better since extraordinarily long words in most documents do not represent words at all, but nonsemantic symbol strings. To optimize space considerations in the database, the choice for maximum word size will usually be adjusted upward. For most natural languages the default maximum word size is usually correct. This option is typically applicable to single-byte European languages and may be ignored by multibyte language processors. (See &cdeman.DtSearch; for more information about DtSearch word sizes). −llang Change the language number to lang. The default is 0. Supported languages include: 0 DtSrLaENG English, ASCII character set 1 DtSrLaENG2 English, ISO Latin-1 character set 2 DtSrLaESP Spanish, ISO Latin-1 character set 3 DtSrLaFRA French, ISO Latin-1 character set 4 DtSrLaITA Italian, ISO Latin-1 character set 5 DtSrLaDEU German, ISO Latin-1 character set 6 DtSrLaJPN Japanese, packed EUC character set; all possible kanji substrings are indexed 7 DtSrLaJPN2 Japanese, packed EUC character set; only individual kanjis are indexed, plus compounds from a knj language file Specifying an unsupported language number will establish a DtSearch custom language for the database. (See &cdeman.DtSearch; for information about DtSearch languages). The dbname operand specifies the new DtSearch database. It consists of an optional path prefix, a 1- to 8-character database name, an optional period, and an optional 1- to 3-character extension. This is the name that the other build tools and the the search API will use to reference the database. None. None. None. The return values are as follows: 0 dtsrcreate completed successfully. non-zero dtsrcreate encountered an error. dtsrcreate reads dtsearch.dbe. It creates or reinitializes the following database files: dbname.d00 dbname.d01 dbname.d21 dbname.d22 dbname.d23 dbname.k00 dbname.k01 dbname.k21 dbname.k22 dbname.k23 It deletes the file dbname.d99. Note that not all necessary database files are created by dtsrcreate. Some additional files are included in the DtSearch distribution, are created by later database build programs, or may be provided by the developer. Create a standard DtSearch type database named mydb that will index ASCII English words of standard length for that language. dtsrcreate mydb Create an AusText type database named jpndb. It will index Japanese words expressed in packed EUC, with automatic compounding of all kanji substrings. When the text contains embedded ASCII, words that are between 2 and 20 characters long will be indexed. At least 150 bytes will be available for the abstract field. dtsrcreate -fa -a150 -wn2 -wx20 -l6 jpndb &cdeman.dtsrdbrec;, &cdeman.DtSrAPI;, &cdeman.dtsrdbfiles;, &cdeman.DtSearch;