DB_File

Ties a Perl hash to one of the Berkeley DB database types and lets you use functions provided in the DB API: 

[$X =] tie %hash, "DB_File", $filename [, $flags, $mode, $DB_HASH]; [$X =] tie %hash, "DB_File", $filename, $flags, $mode, $DB_BTREE; [$X =] tie @array, "DB_File", $filename, $flags, $mode, $DB_RECNO;

The types are:

  • $DB_HASH
  • Stores key/data pairs in data files. Equivalent to other hashing packages such as DBM, NDBM, ODBM, GDBM, and SDBM.
  • $DB_BTREE
  • Stores key/data pairs in a binary tree.
  • $DB_RECNO
  • Uses a record (line) number to access fixed-length and variable-length flat text files through the same key/value-pair interface as in $DB_HASH and $DB_BTREE.

After you've tie d a hash to a database: 

$db = tie %hash, "DB_File", "filename";

you can access the Berkeley DB API functions: 

$db->put($key, $value, R_NOOVERWRITE); # Invoke the DB "put" function

All the functions defined in the dbopen(3) manpage are available except close and dbopen itself. The constants defined in the dbopen manpage are also available.

The following are the functions available (the comments note only the differences from the equivalent C function).

del 
$X->del(key[, flags])

Removes key/value pairs from the database. flags is optional.

fd 
$X->fd

Returns a file descriptor that represents the underlying database. No difference from the equivalent C function.

get 
$X->get(key, value[, flags])

Retrieves data from the database by key. flags is optional. The value associated with key is returned in value.

put 
$X->put(key, value[, flags])

Stores a key/value pair in the database. flags is optional. If R_IAFTER or R_IBEFORE is set, then key is set to the record number of the inserted key/value pair.

seq 
$X->seq(key, value[, flags])

Returns the next sequential key/value pair from the database. flags is optional. Both key and value are set.

sync 
$X->sync([flags])

Synchronizes the database by flushing any cached data to disk. flags is optional.