The RDB operator 'search' may be used to execute one of two fast access methods (binary or hashtable search, although currently only binary search is implemented).
These methods are useful when the key field values of a number of rows in an existing, large rdbtable is known in advance, a common situation.
One example of updating an rdbtable using these methods is as follows. First, 'search' is used to quickly obtain a new, small rdbtable consisting of the desired rows from the existing, large rdbtable. Then 'etbl' is used to update the data in the new rdbtable, including generating new rows, changing some rows, and marking certain rows for deletion, if desired. The next stage would be to use 'mergetbl' to combine the new and old rdbtables into a new, large rdbtable, which will still be in sort order (both the old rdbtable and the new, small one will still exist for backup and/or journaling purposes). Note that the last step could be done in the background.