If this modifier is specified, and the transaction log is not sufficiently large, the import operation will fail. The transaction log must be large enough to accommodate either the number of rows specified by COMMITCOUNT, or the number of rows in the data file if COMMITCOUNT is not specified. It is therefore recommended that the COMMITCOUNT option be specified to avoid transaction log overflow.
This modifier is incompatible with INSERT_UPDATE mode, hierarchical tables, and the USEDEFAULTS modifier.
Replace the fifth paragraph in this section with the following:
When issuing the SET INTEGRITY statement, you can specify the
INCREMENTAL option to explicitly request incremental processing. In most
cases, this option is not needed, because the default behavior is incremental
processing. If incremental processing is not possible, full processing is used
automatically. When the INCREMENTAL option is specified, but incremental
processing is not possible, an error is returned if:
o New constraints have been added to the table or to its parent table while
both tables are in check pending state.
o A load replace operation has taken place, or the NOT LOGGED INITIALLY WITH
EMPTY TABLE option has been activated, after the last integrity check on the table.
o A parent table has been load replaced or checked for integrity non-incrementally.
o The table was in check pending state before migration.
Full processing is required the first time the table is
checked for integrity after migration.
o The table space containing the table or its parent has been
rolled forward to a point in time.
The description of this structure is missing the "indexing_mode" parameter. Supported values (defined in sqlutil.h) are:
SQLU_INX_AUTOSELECT SQLU_INX_REBUILD SQLU_INX_INCREMENTAL SQLU_INX_DEFRRED
For an explanation of these indexing modes, see the description of the LOAD command.
This sample program is missing the "indexing_mode" parameter of the sqluload-in data structure. It should be added as follows:
/* the sqluload input structure */
...
InputInfo.statsopt = SQLU_STATS_NONE; /* don't bother collecting them */
InputInfo.indexing_mode = SQLU_INX_AUTOSELECT; /* let load choose indexing mode */
Change the name of the generic API "db2GenLoadQuery" to "db2gLoadQuery". Change the name of the structure used by the generic API "db2gLoadQuery" to "db2gLoadQueryStruct".
A Fortran version of the "loadqry" sample program is not available.
If SPLIT_FILE_LOCATION points to an NFS directory that is mounted across all partitions after a SPLIT_ONLY operation, a LOAD_ONLY operation will be able to access the split data files without any user intervention. However, if SPLIT_FILE_LOCATION does not point to an NFS directory after a SPLIT_ONLY operation, users are required to manually move the split data files from each partition to the directory in which the LOAD_ONLY operation will be executed.