You can use EDB*Loader with a conventional path load to update the rows within a table, merging new data with the existing data. When you invoke EDB*Loader to perform an update, the server searches the table for an existing row with a matching primary key:To use EDB*Loader to update a table, the table must have a primary key. Please note that you cannot use EDB*Loader to UPDATE a partitioned table.To perform an UPDATE, use the same steps as when performing a conventional path load:
1. Create a data file that contains the rows you wish to UPDATE or INSERT.
2. Define a control file that uses the INFILE keyword to specify the name of the data file. For information about building the EDB*Loader control file, see
3. The following example uses the emp table that is distributed with the Advanced Server sample data. By default, the table contains:The following control file (emp_update.ctl) specifies the fields in the table in a comma-delimited list. The control file performs an UPDATE on the emp table:The data that is being updated or inserted is saved in the emp_update.dat file. emp_update.dat contains:Invoke EDB*Loader, specifying the name of the database (edb), the name of a database superuser (and their associated password) and the name of the control file (emp_update.ctl):edbldr -d edb userid=user_name/password control=emp_update.ctlAfter performing the update, the emp table contains:The rows containing information for the three employees that are currently in the emp table are updated, while rows are added for the new employees (BAKER and MILLS)