Nonstop Database, HiRDB Version 9 UAP Development Guide
![[Contents]](FIGURE/CONTENT.GIF)
![[Index]](FIGURE/INDEX.GIF)
![[Back]](FIGURE/FRONT.GIF)
4.13.3 Notes about performing partial updating and retrieval of BLOB and BINARY data
You should note the following about performing partial updating and retrieval of BLOB and BINARY data:
- A concatenation operation for BLOB or BINARY data can be specified only with an update value in the SET clause of the UPDATE statement. The item specified for the first term in the concatenation operation must be a column specification, and the item specified for the second term must be an embedded variable, the ? parameter, an SQL variable, or an SQL parameter.
For details about the rules for using the concatenation operation to update a BLOB-type or BINARY-type column, see the manual HiRDB Version 9 SQL Reference.
- To execute an addition update, create a column that stores unique key values, and specify that column in the search conditions to identify the update row. To accelerate the row identification process, create an index in that column.
- The minimum input/output unit for BLOB data is the page length of the RDAREA, and HiRDB performs batch input/output of up to 128 kilobytes. Therefore, to improve the performance of BLOB data insertion, addition update, or partial extraction, you should set the data length to units of 128 x 1024 x n bytes (n is a nonzero positive integer).
- If BACKWARD_CUTOFF_UPDATE is not specified in the pd_rpl_func_control operand in the system common definitions, rear deletion update is disabled. In such a case, the data specified in SUBSTR is extracted into memory and then is updated.
All Rights Reserved. Copyright (C) 2011, Hitachi, Ltd.