Following are the parameters that you set in the db2split configuration file:
For delimited data (file type DEL), this parameter is ignored. The record can be any length.
For binary numeric data or packed decimal data (file type BIN or PACK), the exact record length deducted by 1 is used (for backward compatibility), and it has to be less than 32K in length.
For positional ASCII data (file type ASC), and where each record has the same fixed length, specify the actual record length deducted by one. Again, it has to be less than 32K in length.
For positional ASCII data (file type ASC), and where each record has a variable length, and each record is delimited with a line-feed character, it can be set to zero, and the db2split program will distinguish records by the new-line character.
|Note:||The line-feed character in EBCDIC data is X'25'.|
Each record must be delimited by a line-feed character.
All numeric columns in the data file must be in binary format. Supported binary numeric data types include: Integer (4 bytes), Small Integer (2 bytes), Float (4 bytes), Double (8 bytes).
Each record must be the same fixed length and may not be delimited by a new-line character.
All decimal columns in the data file must be in the packed decimal format.
Each record must be the same fixed length.
|Note:||A data file can be any combination of ASC, BIN, or PACK types. The db2split configuration file can contain two declarations of the FileType parameter: one for BIN and one for PACK. If both BIN and PACK are specified in two declarations of the FileType parameter, then all numeric columns must be in binary form, and all decimal columns must be in packed decimal form.|
|Note:||Always use the Mapfili parameter when the partitioning map is customized instead and not the default partitioning map. In the case where the partitioning map is the default, use this parameter or the Mapfili parameter, but not both.|
OutputNodes must be a subset of Nodes.
If this parameter is not specified, output files are created for all database partitions.
The default is stdout.
If OutPutNodes has more than one member, OutPutType is ignored.
Always use this parameter if the partitioning map is customized otherwise only use this parameter if you do not specify Nodes. In an analysis run, you can obtain the set of database partitions that is used to construct the output map from the input partitioning map. An analysis run is established by the RunType parameter described later in this list. Use this parameter or the Nodes parameter but not both to provide a partitioning map to the db2split program.
This parameter is only meaningful during an analysis run.
This file is always written, and can be used as input by the data redistribution utility. The default name is DISTFILE.
If this parameter is defined, all data is written to the specified file. If this parameter is not specified, output from the program is printed to the standard error device. After the filename, specify the mode:
db2split appends a 3-digit suffix (000..999) to the end of the prefix to generate the output file name if the Release parameter is "V5.0". Otherwise, db2split appends a 5-digit suffix (00000...00999) to the end of the prefix to generate the output file name.
The output files are named "prefix suffix". If the OutFile parameter is not specified, the default output filename prefix is NOD.
If FileType was not specified, this parameter can be used to determine if the datafile is ASC or DEL. If this parameter is not specified then the data file is an ASC file; otherwise, the data file is a DEL file. If specified, this parameter can be any character except line-feed, space, binary zero, or carriage-return.
This parameter is only meaningful with DEL files.
By default, the string delimiter is a double quotation mark (") and can be any character except line-feed, space, binary zero, carriage-return, or a period sign (.).
This parameter is only meaningful with DEL files.
By default, the decimal point is a period (.) and can be any character except line-feed, space, carriage-return, or binary zero.
|Note:||CDelimiter, SDelimiter, and DecPt are all
Also, they have to be less than X'40' if the codepage of the data file is a double-byte character set (DBCS), Mixed, or EUC codepage. They cannot be shift-in (SI) or shift-out (SO) characters if the codepage of the data file is EBCDIC Mixed codepage.
Finally, you can specify delimiters in hexidecimal format. For example, you can use X'4F' or 0X'3A'.
|Note:||With CHARACTER, FOR_BIT_DATA, FOR_BIT_VARCHAR, and VARCHAR delimited data, you must specify this field. The length should be equal to the corresponding column length in the database table.|
One Partition statement is used for each column of the partitioning key (from high order to low order). An ASC example is as follows:
If the parameter is YES, header information is generated for all splitting tables. Otherwise, no header information is generated for the splitting files.
The default for this parameter is YES.
The codepage must be a database manager convertible codepage. If it is not provided, it is assigned the codepage number for the database if it is specified; otherwise, it is assigned the codepage number of the application.
The codepage must be a database manager supported codepage. If it is not provided, it is assigned the codepage number of the input data file; otherwise, it is assigned the codepage number of the application.
This parameter is only meaningful for an ASC file.
The parameter values can be YES or NO. The default for this parameter is NO.
If the parameter is YES, and the RecLen parameter is not zero, db2split recognizes records by the new-line delimiter. Then it compares the actual record lengths with the expected record length. If there is no match, an error is returned for that record.
If the parameter is NO, no checking is done.
[ DB2 List of Books | Search the DB2 Books ]