8.3. Using a Command Log File as Input

The procedure for re-executing the commands contained in a Jobname.LOG file or in the database log consists of three steps:

 

8.3.1. Step 1: Establish the Command Log File

The method to do this depends on whether you use the session log file or the database log.

Session Log File

To establish a command log file from the Jobname.log session log file:

  1. List the Jobname.log file by picking Utility Menu> List> Files> Log File.

  2. Pick File> Save as from the Log File window.

Listing the log file in a distributed-memory parallel run displays the log from the master process (Jobname0.log) by default. Use Utility Menu> List> Files> Other... to list the log files on worker processes.

Database Command Log

To establish a command log file from the database log, pick Utility Menu> File> Write DB Log File, or use the LGWRITE command. You can specify a file name or use the default name, Jobname.lgw. You also have the option (with the Kedit field) to write all commands (default), essential commands only (Kedit = REMOVE), or essential commands with nonessential commands commented out (Kedit = COMMENT).

Keep these additional points in mind when using the database command log:

  • The program does not store commands that are read from macros in the database command log. In such cases, you cannot use the LGW file to recreate an exact copy of the database unless all of your macros still exist and are accessible.

  • Files that are read with the /INPUT command will be recorded in the database log (and on Jobname.LOG) only if the LOG field on /INPUT is set to 1.

  • Some "nonessential" commands that are filtered out by the Kedit option on LGWRITE may be required for subsequent *GET operations (for example, EXTREM and PLNSOL). Therefore, use Kedit = COMMENT (not Kedit = REMOVE) to write nonessential commands as comment lines. Using a system editor, you can then "uncomment" the commands required by *GET to ensure exact recreation of the database.

  • When you use the RESUME command to load a previously saved database, the program clears the database log and replaces it with the database log that is stored on the resumed database. (The RESUME command itself is not written to the database command log.)

  • The /CLEAR command (Utility Menu> File> Clear & Start New) clears both the database and the database log. The program stores any commands processed after /CLEAR in the new database that is being created. (The /CLEAR command is not written to the database command log.)

  • When you run in batch mode, the program copies all commands in the input file to the database log before any of these commands execute. If the batch input contains one or more /CLEAR commands, the database log will not match the contents of the database that is produced.

 

8.3.2. Step 2: Edit the Command Log File

Sometimes, you will need to edit your command log file before using it as program input. As you edit your log file, you may want to add comments or indentation to improve its readability. You can add comments to your log file by using comment commands (such as /COM or C***) or by using the comment character (!).

Comments contained in a comment command will appear in both the input and output listings. Comments that follow a ! will appear only in the input listing. You can indent commands to visually group the commands for clarity.

Note:  Certain commands, such as FLST and FITEM, are generated by the GUI and are not intended to be typed in directly. You should avoid editing such commands on the log file. Any data change within the FITEM command, for example, could render the data to be invalid, and could cause unpredictable results.

 

8.3.3. Step 3: Read in the Edited Log File

In an interactive session, pick Utility Menu> File> Read Input from or issue the /INPUT command to read in the edited command log file. In batch mode, you can use the edited command log file as your batch input file.