Nonstop Database, HiRDB Version 9 UAP Development Guide
![[Contents]](FIGURE/CONTENT.GIF)
![[Index]](FIGURE/INDEX.GIF)
![[Back]](FIGURE/FRONT.GIF)
4.12.4 Notes about using the file output facility for BLOB data
- When a BLOB data file that was created becomes unnecessary, the user must delete that file. Note the following point about deleting BLOB data files. Also, BLOB data files can be deleted unconditionally after cursor close or transaction resolution.
- When deleting a BLOB data file immediately after FETCH processing, and the FETCH result prior to the same cursor search and the BLOB value are the same, there are cases in which the file is not re-created with the same file name. In this case, control the processing by storing the prior file name and then deleting it when the file name changes.
- Created BLOB data files are not deleted if an error or a rollback occurs. Note that if the BLOB data files are not deleted, they use up disk space and operating system resources.
- Check that there is enough disk space available before using the following facilities:
- FETCH facility using arrays
Each time FETCH is executed, a file is created for each array element.
- Block transfer facility
During the first FETCH is executed, a file is created for each block transfer row. Subsequently, each time the FETCH for all block transfer rows is completed and the next FETCH is executed, the file creation for each block transfer row is repeated.
- If a file name is the same as the file name of another transaction or cursor search, the files may destroy one another. To avoid this problem, for each transaction or cursor, change the directory or file name in the file prefix so that file names are not duplicated.
All Rights Reserved. Copyright (C) 2011, Hitachi, Ltd.