|HP-UX Reference > N
HP-UX 11i Version 3: February 2007
dbm_open, dbm_close, dbm_fetch, dbm_store, dbm_delete, dbm_firstkey, dbm_nextkey, dbm_error, dbm_clearerr — database subroutines
DBM *dbm_open(const char *file, int flags, mode_t mode);
void dbm_close(DBM *db);
datum dbm_fetch(DBM *db, datum key);
int dbm_store(DBM *db, datum key, datum content, int flags);
int dbm_delete(DBM *db, datum key);
datum dbm_firstkey(DBM *db);
datum dbm_nextkey(DBM *db);
int dbm_error(DBM *db);
int dbm_clearerr(DBM *db);
These functions maintain key/content pairs in a database. They handle very large (a billion blocks (block = 1024 bytes)) databases and can access a keyed item in one or two file system accesses.
key and content parameters are described by the datum type. A datum specifies a string of dsize bytes pointed to by dptr. Arbitrary binary data, as well as normal ASCII strings, are allowed. The database is stored in two files. One file is a directory containing a bit map of keys and has .dir as its suffix. The second file contains all data and has .pag as its suffix.
Before a database can be accessed, it must be opened by dbm_open. This will open and/or create the files file.dir and file.pag depending on the flags parameter (see open(2)).
Once open, the data stored under a key is accessed by dbm_fetch and data is placed under a key by dbm_store. The flags field can be either DBM_INSERT or DBM_REPLACE. DBM_INSERT can only insert new entries into the database, and cannot change an existing entry having the same key. DBM_REPLACE replaces an existing entry if it has the same key. A key (and its associated contents) is deleted by dbm_delete. A linear pass through all keys in a database can be made in (apparently) random order by use of dbm_firstkey and dbm_nextkey. dbm_firstkey returns the first key in the database. dbm_nextkey returns the next key in the database, The following code can be used to traverse the database:
for (key = dbm_firstkey(db); key.dptr != NULL; key = dbm_nextkey(db))
dbm_error returns non-zero when an error has occurred reading or writing the database. dbm_clearerr resets the error condition on the named database.
All functions that return an int indicate errors with negative values and success with zero. Functions that return a datum indicate errors with a null dptr. If dbm_store is called with a flags value of DBM_INSERT and finds an existing entry with the same key, a value of 1 is returned. If a call to dbm_store results in an internal block overflow, a value of -2 is returned.
The ndbm functions provided in this library should not be confused in any way with those of a general-purpose database management system. These functions do not provide for multiple search keys per entry, they do not protect against multi-user access (in other words they do not lock records or files), and they do not provide the many other useful database functions that are found in more robust database management systems. Creating and updating databases by use of these functions is relatively slow because of data copies that occur upon hash collisions. These functions are useful for applications requiring fast lookup of relatively static information that is to be indexed by a single key.
The pointer to data that is returned from these functions are not aligned. This can cause problems if the block contains data that must be aligned to a specific boundry. If the block contains data that must be aligned, the block should be copied to an appropriately aligned area.
The .pag file will contain holes so that its apparent size is about four times its actual content. Some older UNIX systems create real file blocks for these holes when touched. These files cannot be copied by normal means (such as cp(1), cat(1), tar(1), or ar(1)) without expansion.
dptr pointers returned by these subroutines point into static storage that is changed by subsequent calls.
The sum of the sizes of a key/content pair must not exceed the internal block size (currently 1024 bytes). Moreover, all key/content pairs that hash together must fit on a single block. dbm_store returns an error in the event that a disk block fills with inseparable data.
dbm_delete does not physically reclaim file space, although it does make it available for reuse.
The order of keys presented by dbm_firstkey and dbm_nextkey depends on a hashing function, not on anything interesting.
A dbm_store or dbm_delete during a pass through the keys by dbm_firstkey and dbm_nextkey may yield unexpected results.
ndbm(3X) was developed by the University of California, Berkeley.