• The dictionary import/export utility cannot migrate to another HiRDB system the definition information of a table for which an abstract data type is defined. For the information that can be migrated and related details, see 12.1.3 Method that migrates table definition information and table data separately.
• The method that migrates table definition information and table data simultaneously cannot migrate schema units (only table units can be migrated). The method that migrates table definition information and table data separately can migrate schema units and table units.
• To migrate a table that has columns requiring character conversion when the character encoding used differs between the source system and the target system, first convert the character encoding, and then load the data to the target system.
  To migrate a table that has only columns that do not require character conversion (for example, numeric type columns), use the database reorganization utility at the source system to create an input data file for the database load utility, and use the database load utility at the target system to load the data.
  The database reorganization utility causes an error when the character encoding differs between the unload data files of the source system and the target system, but using this method does not cause an error.
  
• A falsification prevented table cannot be moved to another HiRDB system, nor can its table data be reloaded on the target system.
• You cannot use the database copy utility and database recovery utility to migrate data. Do not attempt to migrate data by restoring a backup made on the migration-source system to the migration-destination system.

1. Before migrating table definition information, you must create in the target system an RDAREA that has the same name as the RDAREA in the source system.
2. If you use the method that migrates table definition information and table data simultaneously to migrate a non-partitioned table, even if the RDAREA names differ between the source system and the target system, you can still migrate the table by specifying the -w rdaomit option when you execute the database load utility at the target system. However, if a BLOB column is defined in the table, a table definition error occurs.
3. When using the method that migrates table definition information and table data separately, if the name of the RDAREA that stores the migration target table and the server configuration that stores the RDAREA are different between the source system and the target system, you must specify the -g option when you execute the database reorganization utility. The following figure shows an example of a system configuration in which the -g option must be specified.
   

   Figure 12-1 System configuration in which the -g option must be specified

   

You can change a data type or column definition length only when the database reorganization utility is used to load a file of table data of the source system unloaded in DAT or binary format to the target system. For details about the restrictions on input data files in DAT and binary formats, see the manual HiRDB Version 9 Command Reference.

(a) Cases in which table data cannot be migrated if the table definition is changed

In the following cases, table data cannot be migrated if the table definition is changed:

• Depending on the combination of data types between the source system and the target system, the format (DAT or binary format) of the input data file to be used for migration differs. Consequently, if a single table contains both a data type that can be migrated only in DAT format and a data type that can be migrated only in binary format, you cannot migrate the table. In this case, you must use a UAP or the like to create data for migration.
• When a file in DAT format is migrated, it is converted into a character string during unloading. Therefore, if the file contains data that cannot be converted into a character string, you cannot migrate the table.
  When loading table data to the target system, an error might occur or data might be truncated if the data in the source system does not match the data type in the target system. Therefore, the data in the source system must correspond to the data type in the target system. For details, see the manual HiRDB Version 9 Command Reference.
  

(b) Combinations of data types that can be migrated and input data file formats

The following table shows the combinations of data types that can be migrated between the source system and the target system, and the format of the input data file to create for migration.

Table 12-2 Combinations of data types that can be migrated and input data file formats

Combinations of data types that can be migrated		Input data file format
Source system	Column at the target	DAT	Binary
Numeric data	Numeric data	C	C
	Fixed-size character data	Y	N
	Variable-size character data
	Large object data
	Binary data
Fixed-size character data	Numeric data	C	N
	Fixed-size character data	Y	Y
	Variable-size character data		N
	Date data	C
	Time data
	Date interval data
	Time interval data
	Timestamp data
	Large object data	Y
	Binary data
Variable-size character data	Numeric data	C	N
	Fixed-size character data	Y
	Variable-size character data		Y
	Date data	C	N
	Time data
	Date interval data
	Time interval data
	Timestamp data
	Large object data	Y
	Binary data
Date data	Fixed-size character data	Y	N
	Variable-size character data
	Date data		Y
	Large object data		N
	Binary data
Time data	Fixed-size character data	Y	N
	Variable-size character data
	Date data	N
	Time data	Y	Y
	Large object data		N
	Binary data
Date interval data	Fixed-size character data	Y	N
	Variable-size character data
	Date interval data		Y
	Large object data		N
	Binary data
Time interval data	Fixed-size character data	Y	N
	Variable-size character data
	Time interval data		Y
	Large object data		N
	Binary data
Timestamp data	Fixed-size character data	Y	N
	Variable-size character data
	Timestamp data		Y
	Large object data		N
	Binary data
Large object data	Large object data	N	Y
	Binary data
Binary data	Numeric data	C	N
	Fixed-size character data	Y
	Variable-size character data
	Date data	C
	Time data
	Date interval data
	Time interval data
	Timestamp data
	Large object data
	Binary data	Y	Y

Legend:

Y: Can be migrated.

C: Can be migrated if the data value matches the data type in the target system.

N: Cannot be migrated.

(c) Format of the input data file when the column definition size is changed

The following table shows how definition size is changed in the target system and the format of the input data file to create in each case:

Source system		Target system	Input data file
Data type of the column whose definition size is changed		Definition size after the change (compared with the size before the change)	DAT format	Binary format
Numeric data	DECIMAL type	Greater	C	N
		Smaller	C	N
	Other than DECIMAL type	--#1	--	--
Fixed-size character data		Greater	Y#2	Y#2, #3
		Smaller	C	N
Variable-size character data		Greater	Y	Y
		Smaller	C	N
Timestamp data		Greater	C	N
		Smaller	C	N
Large object data		Greater	N	Y
		Smaller	N	N
Binary data		Greater	C	Y
		Smaller	C	N

Legend:

Y: Can be migrated.

C: Can be migrated if the data value matches the data type in the target system.

N: Cannot be migrated.

--: Not applicable

#1: Numeric data other than DECIMAL-type data is excluded because the definition size is fixed.

#2: The input is left-aligned and the remainder is filled with spaces.

#3: Cannot be migrated if the HiRDB version is earlier than 08-04.