Syntax:
The input file specified after INFILE is scanned. The information about the source table stored there is compared to the definition of the target table specified in the statement.
As long as the only differences are range of value limitations, default values, and index definitions, the structure of the target table is adapted to the structure of the source table during the load prolog. Types and lengths of every column pair formed from source and target table must be identical.
Although source table and target table have the same external structure, they may differ in their internal structure. This is the case when columns have been dropped from one of the tables or columns have been added to an existing table. To be able to continue the load run, the complete column definitions must be loaded in this case.
If a table has indexes that differ from those of the other table, LOAD produces the state of the source file also for the target table; i.e., it drops all the indexes that only exist in the target table and creates the indexes that are available for the source table but are missing in the target table.
The same applies when both tables have different DEFAULT definitions. A column without explicit default, implicitly has the default NULL. The defaults of all columns are passed from the source table to the target table. This means, that explicit defaults of the target table may be deleted.
In interactive mode, the user is asked whether the required adaptations are to be made or cancelled. In case of BATCH runs, LOAD assumes that the user agrees to the adaptations.
The target table must not contain any entries before importing data. This is achieved by using DELETE without qualification.
TABLELOAD, like FASTLOAD, puts the table into READ ONLY mode.