class DBM
Introduction¶ ↑
The DBM
class provides a wrapper to a Unix-style dbm or Database Manager library.
Dbm databases do not have tables or columns; they are simple key-value data stores, like a Ruby Hash
except not resident in RAM. Keys and values must be strings.
The exact library used depends on how Ruby was compiled. It could be any of the following:
-
The original ndbm library is released in 4.3BSD. It is based on dbm library in Unix Version 7 but has different API to support multiple databases in a process.
-
Berkeley DB versions 1 thru 6, also known as BDB and Sleepycat DB, now owned by Oracle Corporation.
-
Berkeley DB 1.x, still found in 4.4BSD derivatives (FreeBSD, OpenBSD, etc).
-
gdbm, the GNU implementation of dbm.
-
qdbm, another open source reimplementation of dbm.
All of these dbm implementations have their own Ruby interfaces available, which provide richer (but varying) APIs.
Cautions¶ ↑
Before you decide to use DBM
, there are some issues you should consider:
-
Each implementation of dbm has its own file format. Generally, dbm libraries will not read each other's files. This makes dbm files a bad choice for data exchange.
-
Even running the same OS and the same dbm implementation, the database file format may depend on the CPU architecture. For example, files may not be portable between PowerPC and 386, or between 32 and 64 bit Linux.
-
Different versions of Berkeley DB use different file formats. A change to the OS may therefore break
DBM
access to existing files. -
Data size limits vary between implementations. Original Berkeley DB was limited to 2GB of data. Dbm libraries also sometimes limit the total size of a key/value pair, and the total size of all the keys that hash to the same value. These limits can be as little as 512 bytes. That said, gdbm and recent versions of Berkeley DB do away with these limits.
Given the above cautions, DBM
is not a good choice for long term storage of important data. It is probably best used as a fast and easy alternative to a Hash
for processing large amounts of data.
Example¶ ↑
require 'dbm' db = DBM.open('rfcs', 0666, DBM::WRCREAT) db['822'] = 'Standard for the Format of ARPA Internet Text Messages' db['1123'] = 'Requirements for Internet Hosts - Application and Support' db['3068'] = 'An Anycast Prefix for 6to4 Relay Routers' puts db['822']
Constants
- NEWDB
Indicates that dbm_open() should open the database in read/write mode, create it if it does not already exist, and delete all contents if it does already exist.
- READER
Indicates that dbm_open() should open the database in read-only mode
- VERSION
Identifies ndbm library version.
Examples:
-
“ndbm (4.3BSD)”
-
“Berkeley DB 4.8.30: (April 9, 2010)”
-
“Berkeley DB (unknown)” (4.4BSD, maybe)
-
“GDBM version 1.8.3. 10/15/2002 (built Jul 1 2011 12:32:45)”
-
“QDBM 1.8.78”
-
- WRCREAT
Indicates that dbm_open() should open the database in read/write mode, and create it if it does not already exist
- WRITER
Indicates that dbm_open() should open the database in read/write mode
Public Class Methods
Open a dbm database with the specified name, which can include a directory path. Any file extensions needed will be supplied automatically by the dbm library. For example, Berkeley DB appends '.db', and GNU gdbm uses two physical files with extensions '.dir' and '.pag'.
The mode should be an integer, as for Unix chmod.
Flags should be one of READER
, WRITER
, WRCREAT
or NEWDB
.
static VALUE fdbm_initialize(int argc, VALUE *argv, VALUE obj) { VALUE file, vmode, vflags; DBM *dbm; struct dbmdata *dbmp; int mode, flags = 0; TypedData_Get_Struct(obj, struct dbmdata, &dbm_type, dbmp); if (rb_scan_args(argc, argv, "12", &file, &vmode, &vflags) == 1) { mode = 0666; /* default value */ } else if (NIL_P(vmode)) { mode = -1; /* return nil if DB not exist */ } else { mode = NUM2INT(vmode); } if (!NIL_P(vflags)) flags = NUM2INT(vflags); FilePathValue(file); /* * Note: * gdbm 1.10 works with O_CLOEXEC. gdbm 1.9.1 silently ignore it. */ #ifndef O_CLOEXEC # define O_CLOEXEC 0 #endif if (flags & RUBY_DBM_RW_BIT) { flags &= ~RUBY_DBM_RW_BIT; dbm = dbm_open(RSTRING_PTR(file), flags|O_CLOEXEC, mode); } else { dbm = 0; if (mode >= 0) { dbm = dbm_open(RSTRING_PTR(file), O_RDWR|O_CREAT|O_CLOEXEC, mode); } if (!dbm) { dbm = dbm_open(RSTRING_PTR(file), O_RDWR|O_CLOEXEC, 0); } if (!dbm) { dbm = dbm_open(RSTRING_PTR(file), O_RDONLY|O_CLOEXEC, 0); } } if (dbm) { /* * History of dbm_pagfno() and dbm_dirfno() in ndbm and its compatibles. * (dbm_pagfno() and dbm_dirfno() is not standardized.) * * 1986: 4.3BSD provides ndbm. * It provides dbm_pagfno() and dbm_dirfno() as macros. * 1991: gdbm-1.5 provides them as functions. * They returns a same descriptor. * (Earlier releases may have the functions too.) * 1991: Net/2 provides Berkeley DB. * It doesn't provide dbm_pagfno() and dbm_dirfno(). * 1992: 4.4BSD Alpha provides Berkeley DB with dbm_dirfno() as a function. * dbm_pagfno() is a macro as DBM_PAGFNO_NOT_AVAILABLE. * 1997: Berkeley DB 2.0 is released by Sleepycat Software, Inc. * It defines dbm_pagfno() and dbm_dirfno() as macros. * 2011: gdbm-1.9 creates a separate dir file. * dbm_pagfno() and dbm_dirfno() returns different descriptors. */ #if defined(HAVE_DBM_PAGFNO) rb_fd_fix_cloexec(dbm_pagfno(dbm)); #endif #if defined(HAVE_DBM_DIRFNO) rb_fd_fix_cloexec(dbm_dirfno(dbm)); #endif #if defined(RUBYDBM_DB_HEADER) && defined(HAVE_TYPE_DBC) /* Disable Berkeley DB error messages such as: * DB->put: attempt to modify a read-only database */ ((DBC*)dbm)->dbp->set_errfile(((DBC*)dbm)->dbp, NULL); #endif } if (!dbm) { if (mode == -1) return Qnil; rb_sys_fail_str(file); } if (dbmp->di_dbm) dbm_close(dbmp->di_dbm); dbmp->di_dbm = dbm; dbmp->di_size = -1; return obj; }
Open a dbm database and yields it if a block is given. See also DBM.new
.
static VALUE fdbm_s_open(int argc, VALUE *argv, VALUE klass) { VALUE obj = fdbm_alloc(klass); if (NIL_P(fdbm_initialize(argc, argv, obj))) { return Qnil; } if (rb_block_given_p()) { return rb_ensure(rb_yield, obj, fdbm_close, obj); } return obj; }
Public Instance Methods
Return a value from the database by locating the key string provided. If the key is not found, returns nil.
static VALUE fdbm_aref(VALUE obj, VALUE keystr) { return fdbm_fetch(obj, keystr, Qnil); }
Stores the specified string value in the database, indexed via the string key provided.
static VALUE fdbm_store(VALUE obj, VALUE keystr, VALUE valstr) { datum key, val; struct dbmdata *dbmp; DBM *dbm; fdbm_modify(obj); keystr = rb_obj_as_string(keystr); valstr = rb_obj_as_string(valstr); key.dptr = RSTRING_PTR(keystr); key.dsize = RSTRING_DSIZE(keystr); val.dptr = RSTRING_PTR(valstr); val.dsize = RSTRING_DSIZE(valstr); GetDBM2(obj, dbmp, dbm); dbmp->di_size = -1; if (dbm_store(dbm, key, val, DBM_REPLACE)) { dbm_clearerr(dbm); if (errno == EPERM) rb_sys_fail(0); rb_raise(rb_eDBMError, "dbm_store failed"); } return valstr; }
Deletes all data from the database.
static VALUE fdbm_clear(VALUE obj) { datum key; struct dbmdata *dbmp; DBM *dbm; fdbm_modify(obj); GetDBM2(obj, dbmp, dbm); dbmp->di_size = -1; while (key = dbm_firstkey(dbm), key.dptr) { if (dbm_delete(dbm, key)) { rb_raise(rb_eDBMError, "dbm_delete failed"); } } dbmp->di_size = 0; return obj; }
Closes the database.
static VALUE fdbm_close(VALUE obj) { struct dbmdata *dbmp; GetDBM(obj, dbmp); dbm_close(dbmp->di_dbm); dbmp->di_dbm = 0; return Qnil; }
Returns true if the database is closed, false otherwise.
static VALUE fdbm_closed(VALUE obj) { struct dbmdata *dbmp; TypedData_Get_Struct(obj, struct dbmdata, &dbm_type, dbmp); if (dbmp->di_dbm == 0) return Qtrue; return Qfalse; }
Deletes an entry from the database.
static VALUE fdbm_delete(VALUE obj, VALUE keystr) { datum key, value; struct dbmdata *dbmp; DBM *dbm; VALUE valstr; long len; fdbm_modify(obj); ExportStringValue(keystr); len = RSTRING_LEN(keystr); if (TOO_LONG(len)) goto not_found; key.dptr = RSTRING_PTR(keystr); key.dsize = (DSIZE_TYPE)len; GetDBM2(obj, dbmp, dbm); value = dbm_fetch(dbm, key); if (value.dptr == 0) { not_found: if (rb_block_given_p()) return rb_yield(keystr); return Qnil; } /* need to save value before dbm_delete() */ valstr = rb_str_new(value.dptr, value.dsize); if (dbm_delete(dbm, key)) { dbmp->di_size = -1; rb_raise(rb_eDBMError, "dbm_delete failed"); } else if (dbmp->di_size >= 0) { dbmp->di_size--; } return valstr; }
Deletes all entries for which the code block returns true. Returns self.
static VALUE fdbm_delete_if(VALUE obj) { datum key, val; struct dbmdata *dbmp; DBM *dbm; VALUE keystr, valstr; VALUE ret, ary = rb_ary_tmp_new(0); int status = 0; long i, n; fdbm_modify(obj); GetDBM2(obj, dbmp, dbm); n = dbmp->di_size; dbmp->di_size = -1; for (key = dbm_firstkey(dbm); key.dptr; key = dbm_nextkey(dbm)) { val = dbm_fetch(dbm, key); keystr = rb_str_new(key.dptr, key.dsize); OBJ_FREEZE(keystr); valstr = rb_str_new(val.dptr, val.dsize); ret = rb_protect(rb_yield, rb_assoc_new(rb_str_dup(keystr), valstr), &status); if (status != 0) break; if (RTEST(ret)) rb_ary_push(ary, keystr); GetDBM2(obj, dbmp, dbm); } for (i = 0; i < RARRAY_LEN(ary); i++) { keystr = RARRAY_AREF(ary, i); key.dptr = RSTRING_PTR(keystr); key.dsize = (DSIZE_TYPE)RSTRING_LEN(keystr); if (dbm_delete(dbm, key)) { rb_raise(rb_eDBMError, "dbm_delete failed"); } } if (status) rb_jump_tag(status); if (n > 0) dbmp->di_size = n - RARRAY_LEN(ary); rb_ary_clear(ary); return obj; }
Calls the block once for each [key, value] pair in the database. Returns self.
static VALUE fdbm_each_pair(VALUE obj) { datum key, val; DBM *dbm; struct dbmdata *dbmp; VALUE keystr, valstr; RETURN_ENUMERATOR(obj, 0, 0); GetDBM2(obj, dbmp, dbm); for (key = dbm_firstkey(dbm); key.dptr; key = dbm_nextkey(dbm)) { val = dbm_fetch(dbm, key); keystr = rb_str_new(key.dptr, key.dsize); valstr = rb_str_new(val.dptr, val.dsize); rb_yield(rb_assoc_new(keystr, valstr)); GetDBM2(obj, dbmp, dbm); } return obj; }
Calls the block once for each key string in the database. Returns self.
static VALUE fdbm_each_key(VALUE obj) { datum key; struct dbmdata *dbmp; DBM *dbm; RETURN_ENUMERATOR(obj, 0, 0); GetDBM2(obj, dbmp, dbm); for (key = dbm_firstkey(dbm); key.dptr; key = dbm_nextkey(dbm)) { rb_yield(rb_str_new(key.dptr, key.dsize)); GetDBM2(obj, dbmp, dbm); } return obj; }
Calls the block once for each [key, value] pair in the database. Returns self.
Calls the block once for each value string in the database. Returns self.
static VALUE fdbm_each_value(VALUE obj) { datum key, val; struct dbmdata *dbmp; DBM *dbm; RETURN_ENUMERATOR(obj, 0, 0); GetDBM2(obj, dbmp, dbm); for (key = dbm_firstkey(dbm); key.dptr; key = dbm_nextkey(dbm)) { val = dbm_fetch(dbm, key); rb_yield(rb_str_new(val.dptr, val.dsize)); GetDBM2(obj, dbmp, dbm); } return obj; }
Returns true if the database is empty, false otherwise.
static VALUE fdbm_empty_p(VALUE obj) { datum key; struct dbmdata *dbmp; DBM *dbm; GetDBM2(obj, dbmp, dbm); if (dbmp->di_size < 0) { dbm = dbmp->di_dbm; for (key = dbm_firstkey(dbm); key.dptr; key = dbm_nextkey(dbm)) { return Qfalse; } } else { if (dbmp->di_size) return Qfalse; } return Qtrue; }
Return a value from the database by locating the key string provided. If the key is not found, returns ifnone
. If ifnone
is not given, raises IndexError
.
static VALUE fdbm_fetch_m(int argc, VALUE *argv, VALUE obj) { VALUE keystr, valstr, ifnone; rb_scan_args(argc, argv, "11", &keystr, &ifnone); valstr = fdbm_fetch(obj, keystr, ifnone); if (argc == 1 && !rb_block_given_p() && NIL_P(valstr)) rb_raise(rb_eIndexError, "key not found"); return valstr; }
Returns true if the database contains the specified key, false otherwise.
Returns true if the database contains the specified string value, false otherwise.
static VALUE fdbm_has_value(VALUE obj, VALUE valstr) { datum key, val; struct dbmdata *dbmp; DBM *dbm; long len; ExportStringValue(valstr); len = RSTRING_LEN(valstr); if (TOO_LONG(len)) return Qfalse; val.dptr = RSTRING_PTR(valstr); val.dsize = (DSIZE_TYPE)len; GetDBM2(obj, dbmp, dbm); for (key = dbm_firstkey(dbm); key.dptr; key = dbm_nextkey(dbm)) { val = dbm_fetch(dbm, key); if ((DSIZE_TYPE)val.dsize == (DSIZE_TYPE)RSTRING_LEN(valstr) && memcmp(val.dptr, RSTRING_PTR(valstr), val.dsize) == 0) return Qtrue; } return Qfalse; }
Returns true if the database contains the specified key, false otherwise.
static VALUE fdbm_has_key(VALUE obj, VALUE keystr) { datum key, val; struct dbmdata *dbmp; DBM *dbm; long len; ExportStringValue(keystr); len = RSTRING_LEN(keystr); if (TOO_LONG(len)) return Qfalse; key.dptr = RSTRING_PTR(keystr); key.dsize = (DSIZE_TYPE)len; GetDBM2(obj, dbmp, dbm); val = dbm_fetch(dbm, key); if (val.dptr) return Qtrue; return Qfalse; }
Returns a Hash
(not a DBM
database) created by using each value in the database as a key, with the corresponding key as its value.
static VALUE fdbm_invert(VALUE obj) { datum key, val; struct dbmdata *dbmp; DBM *dbm; VALUE keystr, valstr; VALUE hash = rb_hash_new(); GetDBM2(obj, dbmp, dbm); for (key = dbm_firstkey(dbm); key.dptr; key = dbm_nextkey(dbm)) { val = dbm_fetch(dbm, key); keystr = rb_str_new(key.dptr, key.dsize); valstr = rb_str_new(val.dptr, val.dsize); rb_hash_aset(hash, valstr, keystr); } return hash; }
Returns the key for the specified value.
static VALUE fdbm_key(VALUE obj, VALUE valstr) { datum key, val; struct dbmdata *dbmp; DBM *dbm; long len; ExportStringValue(valstr); len = RSTRING_LEN(valstr); if (TOO_LONG(len)) return Qnil; GetDBM2(obj, dbmp, dbm); for (key = dbm_firstkey(dbm); key.dptr; key = dbm_nextkey(dbm)) { val = dbm_fetch(dbm, key); if ((long)val.dsize == RSTRING_LEN(valstr) && memcmp(val.dptr, RSTRING_PTR(valstr), val.dsize) == 0) { return rb_str_new(key.dptr, key.dsize); } } return Qnil; }
Returns true if the database contains the specified key, false otherwise.
Returns an array of all the string keys in the database.
static VALUE fdbm_keys(VALUE obj) { datum key; struct dbmdata *dbmp; DBM *dbm; VALUE ary; GetDBM2(obj, dbmp, dbm); ary = rb_ary_new(); for (key = dbm_firstkey(dbm); key.dptr; key = dbm_nextkey(dbm)) { rb_ary_push(ary, rb_str_new(key.dptr, key.dsize)); } return ary; }
Returns the number of entries in the database.
static VALUE fdbm_length(VALUE obj) { datum key; struct dbmdata *dbmp; DBM *dbm; int i = 0; GetDBM2(obj, dbmp, dbm); if (dbmp->di_size > 0) return INT2FIX(dbmp->di_size); for (key = dbm_firstkey(dbm); key.dptr; key = dbm_nextkey(dbm)) { i++; } dbmp->di_size = i; return INT2FIX(i); }
Returns true if the database contains the specified key, false otherwise.
Converts the contents of the database to an in-memory Hash
, then calls Hash#reject
with the specified code block, returning a new Hash
.
static VALUE fdbm_reject(VALUE obj) { return rb_hash_delete_if(fdbm_to_hash(obj)); }
Deletes all entries for which the code block returns true. Returns self.
Replaces the contents of the database with the contents of the specified object. Takes any object which implements the each_pair
method, including Hash
and DBM
objects.
static VALUE fdbm_replace(VALUE obj, VALUE other) { fdbm_clear(obj); rb_block_call(other, rb_intern("each_pair"), 0, 0, update_i, obj); return obj; }
Returns a new array consisting of the [key, value] pairs for which the code block returns true.
static VALUE fdbm_select(VALUE obj) { VALUE new = rb_ary_new(); datum key, val; DBM *dbm; struct dbmdata *dbmp; GetDBM2(obj, dbmp, dbm); for (key = dbm_firstkey(dbm); key.dptr; key = dbm_nextkey(dbm)) { VALUE assoc, v; val = dbm_fetch(dbm, key); assoc = rb_assoc_new(rb_str_new(key.dptr, key.dsize), rb_str_new(val.dptr, val.dsize)); v = rb_yield(assoc); if (RTEST(v)) { rb_ary_push(new, assoc); } GetDBM2(obj, dbmp, dbm); } return new; }
Removes a [key, value] pair from the database, and returns it. If the database is empty, returns nil. The order in which values are removed/returned is not guaranteed.
static VALUE fdbm_shift(VALUE obj) { datum key, val; struct dbmdata *dbmp; DBM *dbm; VALUE keystr, valstr; fdbm_modify(obj); GetDBM2(obj, dbmp, dbm); dbmp->di_size = -1; key = dbm_firstkey(dbm); if (!key.dptr) return Qnil; val = dbm_fetch(dbm, key); keystr = rb_str_new(key.dptr, key.dsize); valstr = rb_str_new(val.dptr, val.dsize); dbm_delete(dbm, key); return rb_assoc_new(keystr, valstr); }
Stores the specified string value in the database, indexed via the string key provided.
Converts the contents of the database to an array of [key, value] arrays, and returns it.
static VALUE fdbm_to_a(VALUE obj) { datum key, val; struct dbmdata *dbmp; DBM *dbm; VALUE ary; GetDBM2(obj, dbmp, dbm); ary = rb_ary_new(); for (key = dbm_firstkey(dbm); key.dptr; key = dbm_nextkey(dbm)) { val = dbm_fetch(dbm, key); rb_ary_push(ary, rb_assoc_new(rb_str_new(key.dptr, key.dsize), rb_str_new(val.dptr, val.dsize))); } return ary; }
Converts the contents of the database to an in-memory Hash
object, and returns it.
static VALUE fdbm_to_hash(VALUE obj) { datum key, val; struct dbmdata *dbmp; DBM *dbm; VALUE hash; GetDBM2(obj, dbmp, dbm); hash = rb_hash_new(); for (key = dbm_firstkey(dbm); key.dptr; key = dbm_nextkey(dbm)) { val = dbm_fetch(dbm, key); rb_hash_aset(hash, rb_str_new(key.dptr, key.dsize), rb_str_new(val.dptr, val.dsize)); } return hash; }
Returns true if the database contains the specified string value, false otherwise.
Returns an array of all the string values in the database.
static VALUE fdbm_values(VALUE obj) { datum key, val; struct dbmdata *dbmp; DBM *dbm; VALUE ary; GetDBM2(obj, dbmp, dbm); ary = rb_ary_new(); for (key = dbm_firstkey(dbm); key.dptr; key = dbm_nextkey(dbm)) { val = dbm_fetch(dbm, key); rb_ary_push(ary, rb_str_new(val.dptr, val.dsize)); } return ary; }
Returns an array containing the values associated with the given keys.
static VALUE fdbm_values_at(int argc, VALUE *argv, VALUE obj) { VALUE new = rb_ary_new2(argc); int i; for (i=0; i<argc; i++) { rb_ary_push(new, fdbm_fetch(obj, argv[i], Qnil)); } return new; }