TurboIMAGE/XL Database Management System Reference Manual (30391-90012)
Table Of Contents
- 1 Introduction
- 2 Database Structure and Protection
- 3 Defining a Database
- 4 Using the Database
- 5 TurboIMAGE/XL Library Procedures
- Using TurboIMAGE/XL Intrinsics
- DBBEGIN
- DBCLOSE
- DBCONTROL
- DBDELETE
- DBEND
- DBERROR
- DBEXPLAIN
- DBFIND
- DBGET
- DBINFO
- INTRINSIC NUMBER 402
- Syntax
- Parameters
- Discussion
- Mode 101: Item Number
- Mode 102: Item Name
- Mode 103: Items in Database
- Mode 104: Items in Data Set
- Mode 113: BTREEMODE1 and Wildcard Character
- Mode 201: Set Number
- Mode 202: Set Name
- Mode 203: Sets in Database
- Mode 204: Sets with Item
- Mode 205: Set Capacity
- Mode 206: Number of Data Set Chunks
- Mode 207: Size of Data Set Chunks
- Mode 208: Primary and Actual Capacity
- Mode 209: B-Tree Attachment
- Mode 301: Paths
- Mode 302: Key or Search Item
- Mode 401: Logging
- Mode 402: ILR
- Mode 403: Dynamic Roll-Back
- Mode 404: Logging Subsystem Information
- Mode 406: Database Information
- Mode 501: Subsystem Access
- Mode 502: Critical Item Update
- Modes 8nn: Third-Party Indexing
- Mode 901: Language
- DBLOCK
- DBMEMO
- DBOPEN
- DBPUT
- DBUNLOCK
- DBUPDATE
- DBXBEGIN
- DBXEND
- DBXUNDO
- 6 Host Language Access
- Model Program
- ORDERS Database Model Program
- Main Body of Program
- Opening the Database
- Retrieving All the Records on a Chain (with Item Level Locking)
- Retrieving a Data Entry Using a Record Number
- Retrieving Master Data Using a Key Value
- Retrieving Data Serially (with Set Level Locking)
- Adding an Entry
- Updating an Entry
- Deleting an Entry
- Rewinding a Data Set
- Obtaining Database Information
- Obtaining Error Messages and Explanations
- Closing the Database
- C
- COBOL II
- Defining Data Types, Variables, and Intrinsics
- Main Body of Program
- Opening the Database
- Retrieving All the Records on a Chain (with Item Level Locking)
- Retrieving a Data Entry Using a Record Number
- Retrieving Master Data Using a Key Value
- Retrieving Data Serially (with Set Level Locking)
- Adding an Entry
- Updating an Entry
- Deleting an Entry
- Rewinding a Data Set
- Obtaining Database Information
- Obtaining Error Messages and Explanations
- Closing the Database
- FORTRAN 77
- Pascal
- RPG
- 7 Logging and Recovery
- Database Utilities Used in Logging and Recovery
- Recovery Options
- Logical Transactions
- Dynamic Roll-Back Recovery
- Intrinsic Level Recovery
- Logging Preparation
- Step 1--Checking MPE/iX Logging Configuration
- Step 2--Acquiring Logging Capability
- Step 3--Logging to Tape or Disk
- Step 4--Building a Log File for Logging to Disk
- Step 5--Creating the Log Identifier
- Step 6--Setting the Log Identifier
- Step 7--Setting Flags for the Database Backup Copy
- Step 8--Making a Database Backup Copy
- TurboSTORE/iX 7x24 True-Online Backup
- Logging Status
- Logging Maintenance
- Roll-Forward Recovery
- Roll-Back Recovery
- DBRECOV Commands Used with Roll-Forward and Roll-Back Recovery
- Recovery Tables
- Post-Recovery Options
- The Mirror Database
- 8 Using the Database Utilities
- Restructuring the Database with TurboIMAGE/XL Utilities
- Summary of Utility Routines
- Utility Program Operation
- DBLOAD
- DBRECOV
- >CONTROL
- >EXIT
- >FILE
- >RECOVER
- >ROLLBACK
- >RUN
- DBRESTOR
- DBSTORE
- DBUNLOAD
- DBUTIL
- >>ACTIVATE
- >>ADDINDEX
- >>CREATE
- >>DEACTIVATE
- >>DETACH
- >>DISABLE
- >>DROPINDEX
- >>ENABLE
- >>ERASE
- >>EXIT
- >>HELP
- >>MOVE
- >>PURGE
- >>REBUILDINDEX
- >>REDO
- >>RELEASE
- >>SECURE
- >>SET
- >>SHOW
- Syntax
- Parameters
- Example (Show Users)
- Example Discussion
- Example (Show All)
- Example Discussion
- Example (Show Capacity)
- Format of Show Device List
- Example (Show Device)
- Format of Show Indices
- Example (Show Indices)
- Format of Show Locks List
- Example 1 (Show Locks)
- Example 1 Discussion
- Example 2 (Show Locks)
- Example 2 Discussion
- >>VERIFY
- 9 Using a Remote Database
- 10 Internal Structures and Techniques
- 11 B-Tree Indices
- A Error Messages
- B Results of Multiple Access
- C Database Design Considerations
- D Multiple Calls to DBLOCK
- E TurboIMAGE/XL Log Record Formats
- F MPE/iX Log Record Formats
- G Recovery and Logging Quick Reference
- H TurboIMAGE/XL versus TurboIMAGE/V

Chapter 10 495
Internal Structures and Techniques
Internal Techniques
before the record is added, the assigned record address is the primary address. If the bit is
a one before the record is added, it indicates that an entry already exists. If this existing
entry is a primary entry, a search is done to find a free location, secondary address, for the
new entry. However, if the existing entry is a secondary entry, this secondary entry is
relocated to another free secondary address, and the new entry is added at this location. If
the data set is enabled for dynamic expansion, the search for a free secondary address is
done in the original as well as expansion area. For the original (primary, hashing, or
initial) area, a secondary address is identified by a serial search of the bit maps of blocks
for a zero indicating an unused record. For the expanded area, a secondary address is
identified using the pointer to a delete chain and end-of-file pointer, as in detail data sets.
Space Allocation for Detail Data Sets
Space allocation for each detail data set is controlled by a free space counter, an end-of-file
pointer and a pointer to a delete chain. The end-of-file pointer contains the record
address of the highest-numbered entry which has existed so far in the data set. The delete
chain pointer contains the record address of the entry which was most recently deleted.
When each detail data set is first created, the end-of-file pointer and delete chain pointer
are both zero.
When a new entry is added to a detail data set, TurboIMAGE/XL assigns to it the record
address referenced by the delete chain pointer, unless the pointer is zero or HWMPUT has
been enabled either using DBCONTROL or using DBUTIL. If the delete chain pointer is zero or
if the HWMPUT flag is enabled, the end-of-file pointer is incremented and then used as the
assigned record address. The free space counter is decremented in either case. When a new
entry is to be added to a detail data set and the free space counter is zero, at run-time, the
data set is expanded according to the capacity expansion parameter specified for this data
set (if any). The capacity expansion parameters can be set using DBSCHEMA for new
databases or by using DBCHANGE Plus or other third-party utilities for existing databases.
When an existing entry is deleted, its media record is zeroed, the first word is replaced
with the current delete chain pointer, and the block is written to disk. The delete chain
pointer is set to the address of the newly deleted entry and the free space counter is
incremented.
The delete chain is, in effect, a "last-in–first-out" linked list of reusable media record space.
Reusable space is allocated in preference to the unused space represented by record
addresses beyond the end-of-file pointer except when HWMPUT is enabled.
Addition and deletion of data entries also requires data chain maintenance and turning on
or turning off the corresponding bit of the appropriate bit map. Both of these are necessary
for retrieval integrity but neither play a role in space allocation for detail data sets.