DATABASE FILES

The basic type of files which Star4Win generates and works with are database files, with a usual extension .dbf. These files are generally compatible with widely used software products like dBASE III+, dBASE IV and FoxBase. Star4Win database files, however, have their own peculiar features. In order to explain them, we have to dwell in some detail on databases in general.

A database file (sometimes called electronic table or a spreadsheet) is a table-like structure consisting of a certain amount of records ( = lines), each record being subdivided into fields ( = columns). The number, types and names of fields are fixed and can be changed only by means of a special procedure (called "modifying structure"), whereas the number of records is optional and can be easily increased or decreased by the user. The fixed structure of the database files allows for a very quick access to all records even in huge databases.

Fields in a database file can belong to several types, serving different purposes. Standard dBase files have five basic field types:

Character fields

For input of all kinds of character information.

Numeric fields

For numbers only.

Date fields

For dates only.

Logical fields

For binary logical input (these fields may only contain the information "Yes"/"No" or "True/False")

Memo fields

For input of character (memo) information.

Memo fields in standard dBase are, in fact, not fields in the strict sense of the word. The database itself contains only numeric pointers to the field information contained in a separate file (with a standard extension .dbt). These numeric pointers themselves occupy special fields (10 bytes long for every memo field) within the .dbf file. Once you input anything (even 1 symbol) into a memo field within a certain record, a standard block of 512 bytes is added to the .dbt file - and most of these 512 bytes remain unused. On the other hand, if you edit an already occupied memo field and cross the 512-byte boundary, the whole content of the field is overwritten in two 512-byte blocks which are then added to the end of the .dbt file - while the old 512- byte block stays in its original place (although it is not being used any more). All this results in the fact that .dbt files may grow even longer than .dbf files.

Star4Win supports all the original dBase field types, which means that you can safely view and edit original dBase files while working with Star4Win. It has, however, its own specific features:

• Character fields are no longer restricted in length (the maximum field length is 65535 symbols).
• The number of possible fields is increased (from a dBase III+ maximum in 127 fields to 2046 fields).
• Star4Win also introduces a new type of fields: fields with variable length. They can be effectively used for any character information and usually occupy much less space than standard character fields and memo fields (which are still supported for compatibility).
      Variable length fields are in many respects similar to memo fields: each variable ("V") field occupies 6 bytes within a database file, and their contents are actually stored in a separate file with the extension .var. However, each entry within a V-field occupies exactly the space it needs.
      You should remember several things concerning the new field type:
      a) Special Star4Win files (wordlists etc.) are created without V-fields. Any character (or memo) field, however, can be easily transformed to a V-field by using the File structure option of the File menu.
      b) You can create new files (using the Create a database menu option) already with V-fields, or else modify the structure of already existing files.
      c) All operations that were supported for character ("C") and memo ("M") fields, are also supported for V- fields.
      d) In Star4Win, when you edit your database files, contents of each field (whether it is a C-field, M-field or Vfield) immediately appear on the screen. However, there may be cases when long M-fields or V-fields do not fit into one screen. In such cases you can press Ctrl T to enter the full-screen editing mode, and the content of the field will be edited just like a text file with a text editor. If you wish to save the field content as it is (with margins, line breaks and indents), please enter the  character (by pressing Ctrl and then [) in the field's beginning.
      e) The .var files (containing the variable length fields) also have a tendency to grow unnecessarily, although not at all as rapidly as the .dbt files with memo fields. This happens when you make a new input within an already occupied record, overwriting its previous content and exceeding it in length. In this case the new field content is written to the end of file, while the old record stays in its original place. You may shrink your files periodically, choosing the option Pack in the Assist menu. This command first does the checking of the variable fields fragmentation: if it is encountered, you are asked whether you wish to pack your file or not. Only afterwards the usual routine of packing deleted records (if any) is performed.
      Periodical PACKing of your files is also recommended, because some operations (e.g., searching for string values using the F3 key) may be hindered by the file's fragmentation.
  
• Starting with version 1.0.7 the database system of StarLing / Star4Win has been enhanced by field aliases. A field alias is a field description that can be optionally used instead of the real field name in database headers. Unlike field names, field aliases are not restricted in length, can consist of several words, contain spaces and non-Roman (e.g. Cyrillic) letters. You can change field aliases while modifying structure or while changing any field name (Change field name... option of the Field menu).
     A single field may have multiple aliases: in this case they should be preceded by slash + number + semicolon (e.g. \1:Fieldname1\2:Fieldname2). All aliases having the same number are united within an alias set, and it is possible to assign any of these sets to your database file. Note that if the alias set 0 is chosen, only original field names will be displayed. 
      Field aliases, as well as some other information, are stored in special text files with the extension .inf (officially also introduced in version 1.0.7). These files are optional - i.e., the .dbf/.var files will still be editable without .inf files, but many useful parameters can be stored in this way. The parameters that .inf files now support are:
  • a) Links to related databases. E.g., in the database \data\alt\turcet containing the field PROTO the line PROTO = \data\alt\altet means that the field PROTO relates turcet.dbf to altet.dbf.
              In many cases (for standard lexicostatistical files) there exist automatic relations to lower level and upper level files; note that the settings within INF files override automatic relations.
  • b) field aliases and alias sets (see below): the syntax is FIELD_ALIAS[field name] = alias; ALIASSET = 1 (2, 3...n)
  • c) default indexes (see below): the syntax is INDEX = index file name
  • d) list of fields to display and order of fields. Here the syntax is: FIELD_LIST = field range
  • e) list (of unlimited length) of databases that can be linked to any segment of text surrounded by the delimiters \X0....\x, \X1....\x etc. Every database name should be located on a new line with the syntax LOOKUP[1] = database file) etc. Links of this type differ from global links specified in config.str in that the list of linked databases may be different for any individual database, thus allowing for more flexibility. At this time for such links you cannot specify default reference fields; however, it is possible to directly specify record numbers in linked databases, see below on links.
    Note: \X0....\x refers to the current database.
  • f) file description - a text of unlimited length that may contain any information concerning the current database. There can be any number of file descriptions, corresponding to the number of alias sets available for the file. The descriptions should be started by the line DBINFO START number and ended by the line DBINFO END.
  You see that .inf files have become a rather important part of the database system - if you wish to skip the necessity of readjusting the list of linked databases, displayed fields and their order, as well as the field aliases every time you open the database you are working it. So if you transfer a database to a different directory, drive or computer, do not forget to copy the .inf file and place it in the same folder.
      All the parameters within the .inf file can be viewed or edited by the Star32 text editor, or in fact by any text editor program. However, for user convenience there are now several new options in the File menu specially suited for this purpose. Those are:
  • File info: general technical information about the current database (this option was there before but was located in the Assist menu).
  • File configuration: A menu dialog similar to the general configuration, but allowing to edit information pertaining to the current file. So far it contains two sub-options:
    
    • INF file content: here you can view and edit the .inf file.
      For user convenience the options of editing the .inf file as text and editing file decriptions are also located in the general File menu, as INF file text edit and File description, with the keyboard shortcuts Alt F2 and Alt F1 respectively.
    • Lookup databases: here you can add any number of lookup databases for hyperlinks within your file.