Introduction to File and File Index Management
Reality file types
Reality supports the following types of file:
-
Normal. An ordinary Reality file on the local database. A normal file has the basic structure described in File and Item Structures. Normal files are backed up via the usual Reality save and restore procedures.
-
Foreign Database. This is a Reality file which has nearly all the characteristics of a normal file, with the structure described in File and Item Structures, but which is located on a foreign (SQL-based) database. You can create, update and manage Foreign Database files in the same way as Normal files. Multi-values and binary items are supported. The files also support indexing and Transaction Handling.
It is unlikely that you would want to back up Foreign Database files using the Reality save commands. If a Reality application is storing its data on a foreign database, it is recommended that this data is secured by saving the SQL database in whatever way is appropriate.
-
SQL View. This is a Reality file that provides a view of an existing SQL table (or SQL view) on a foreign database. A table, or view, is mapped to the Reality file; each row in that table becomes an item in the Reality file and each column within that row is an attribute in the item.
As the data is stored in a form compatible with the foreign database's native applications, restrictions are imposed on Reality applications writing such data. There is no support for multi-values, indexing or Transaction Handling.
SQL-view files are primarily used for read access, but it is possible to update the external table data if you know the exact structure and controls. If you save an SQL view file, you are saving the file definition item, but not the external data, which is secured on the foreign database.
-
Directory View. A Reality file that provides a view of a directory in the host system (UNIX or Windows). A directory is mapped to the Reality file; each file in the directory appears as an item in that file.
Directory view files are primarily aimed at text file manipulation. A directory view file contains Reality items for regular system text files within the referenced directory. Non-regular files - such as sub-directories, pipes and devices - are not visible as items.
-
Hyper File. A hyper file is an updatable view onto a number of discrete data sections. Accessing a hyper file accesses its constituent data sections as if they were one file, and their indexes as if they were one index.
Creating a file
The following commands enable you to create files on your database:
-
Creates a normal or Foreign Database file.
-
Creates a file that provides a view of a directory in the host system (UNIX or Windows, as appropriate) from within Reality.
-
Creates a file that provides a view of an SQL table or SQL view from within Reality.
-
Creates a hyper file.
Deleting files and items
The following commands allow you to remove files and items from your database:
-
Deletes all items, except data descriptors, from a specified file.
-
Delete selected items from specified file. EDELETE is an English command.
-
Deletes a specified file, directory view, SQL view or hyper file from the current account.
-
Removes a specific section, or all sections, from a hyper file by updating the lookup table in the hyper file definition item (H-pointer) and, optionally, deletes the sections themselves.
Copying an item
The following commands allow you to copy selected items from a file to another file, terminal or printer:
-
Copies selected items from a file to another file, terminal or printer.
-
Copies selected items from a file to a printer.
-
Copies selected items from a file to a terminal's screen.
-
Dumps contents of selected items to terminal screen or to printer, displaying system delimiters. These are English commands.
Reformatting data
The commands REFORMAT and SREFORMAT can be used to select and re-order attributes of selected file items and write them to a new file. These are English commands.
Searching for an item-id
The FIND command allows you to search all files in your current account for a specified item-id.
Searching for a string
The SEARCH and ESEARCH commands allow you to search for a specified string within a particular file. ESEARCH is an English command.
Listing files in an account
A number of different commands allow you to list files in your current account according to particular selection criteria and display them on your screen:
-
Lists all dictionaries in the current account.
-
Lists all dictionaries and data sections in an account or specified dictionary.
-
Displays all Q-pointers in an account MD.
-
Display all files in the current account in two-, three- or four-column format.
Listing and sorting items in a file
The following commands allow you to display on your terminal a list of item-ids in a specified file:
-
Lists the item-ids of all items in a specified file in four-column format.
-
Sort all item-ids in a specified file in ASCII character order and list them in two-, three- or four-column format.
Refer to the English Reference for ways of reporting the contents of a file in more detail.
Moving and renaming files
These commands are provided to move and/or rename files:
-
Moves a file from one account to another, or a data section from one file to another, and optionally renames the moved file/data section. Also used to rename an account.
-
Moves items from one file/data section to another file or to another data section of the same file, while preserving their time-stamps.
-
Renames an entire file or a file data section.
File hashing statistics
The following commands display file hashing statistics and optionally a histogram:
-
Displays the distribution of items that would result from a different modulo.
-
Displays information about the current distribution of items in the file.
-
Displays a series of prime numbers that can be used as modulos, from one to greater than 16 million. Each prime number differs from the previous one by at least 10%.
Comparing items
The COMP and COMPARE commands allow you to perform a comparison between items in the same file, or in two different files, and generate a report.
Managing indexes
The following commands are used to manage file indexes:
-
Creates a index definition item in the file dictionary which contains the selection and sort criteria used to create an index. The definition item-id is the index name. It can also be used to recover a deleted or corrupt index definition from an existing index.
-
Creates an index section for a particular data section, based on a specified index definition item.
-
Deletes an index section previously generated by CREATE-INDEX.
-
Generates an active list of selected item-ids in an index, sorted in ascending sequence. The list can then be used by an immediately following English or TCL-II command, including SAVE-LIST.
-
Checks an index against its data section for non-existent item references, or key values that do not match the index sort criteria. Errors must be corrected by deleting the index section using DELETE-INDEX and recreating it using CREATE-INDEX.
-
Lists the indexes for a particular file, data section or account. The indexes are verified and any errors reported.
Optimising file size (partition database only)
Two TCL commands, SIZE-MONITOR and RESIZE-FILE, allow administrators to maintain optimally sized files on a partition database.