2 * @brief Lightning memory-mapped database library
4 * @mainpage Lightning Memory-Mapped Database Manager (LMDB)
6 * @section intro_sec Introduction
7 * LMDB is a Btree-based database management library modeled loosely on the
8 * BerkeleyDB API, but much simplified. The entire database is exposed
9 * in a memory map, and all data fetches return data directly
10 * from the mapped memory, so no malloc's or memcpy's occur during
11 * data fetches. As such, the library is extremely simple because it
12 * requires no page caching layer of its own, and it is extremely high
13 * performance and memory-efficient. It is also fully transactional with
14 * full ACID semantics, and when the memory map is read-only, the
15 * database integrity cannot be corrupted by stray pointer writes from
18 * The library is fully thread-aware and supports concurrent read/write
19 * access from multiple processes and threads. Data pages use a copy-on-
20 * write strategy so no active data pages are ever overwritten, which
21 * also provides resistance to corruption and eliminates the need of any
22 * special recovery procedures after a system crash. Writes are fully
23 * serialized; only one write transaction may be active at a time, which
24 * guarantees that writers can never deadlock. The database structure is
25 * multi-versioned so readers run with no locks; writers cannot block
26 * readers, and readers don't block writers.
28 * Unlike other well-known database mechanisms which use either write-ahead
29 * transaction logs or append-only data writes, LMDB requires no maintenance
30 * during operation. Both write-ahead loggers and append-only databases
31 * require periodic checkpointing and/or compaction of their log or database
32 * files otherwise they grow without bound. LMDB tracks free pages within
33 * the database and re-uses them for new write operations, so the database
34 * size does not grow without bound in normal use.
36 * The memory map can be used as a read-only or read-write map. It is
37 * read-only by default as this provides total immunity to corruption.
38 * Using read-write mode offers much higher write performance, but adds
39 * the possibility for stray application writes thru pointers to silently
40 * corrupt the database. Of course if your application code is known to
41 * be bug-free (...) then this is not an issue.
43 * If this is your first time using a transactional embedded key/value
44 * store, you may find the \ref starting page to be helpful.
46 * @section caveats_sec Caveats
47 * Troubleshooting the lock file, plus semaphores on BSD systems:
49 * - A broken lockfile can cause sync issues.
50 * Stale reader transactions left behind by an aborted program
51 * cause further writes to grow the database quickly, and
52 * stale locks can block further operation.
54 * Fix: Check for stale readers periodically, using the
55 * #mdb_reader_check function or the \ref mdb_stat_1 "mdb_stat" tool.
56 * Stale writers will be cleared automatically on most systems:
57 * - Windows - automatic
58 * - BSD, systems using SysV semaphores - automatic
59 * - Linux, systems using POSIX mutexes with Robust option - automatic
60 * Otherwise just make all programs using the database close it;
61 * the lockfile is always reset on first open of the environment.
63 * - On BSD systems or others configured with MDB_USE_SYSV_SEM or
65 * startup can fail due to semaphores owned by another userid.
67 * Fix: Open and close the database as the user which owns the
68 * semaphores (likely last user) or as root, while no other
69 * process is using the database.
71 * Restrictions/caveats (in addition to those listed for some functions):
73 * - Only the database owner should normally use the database on
74 * BSD systems or when otherwise configured with MDB_USE_POSIX_SEM.
75 * Multiple users can cause startup to fail later, as noted above.
77 * - There is normally no pure read-only mode, since readers need write
78 * access to locks and lock file. Exceptions: On read-only filesystems
79 * or with the #MDB_NOLOCK flag described under #mdb_env_open().
81 * - An LMDB configuration will often reserve considerable \b unused
82 * memory address space and maybe file size for future growth.
83 * This does not use actual memory or disk space, but users may need
84 * to understand the difference so they won't be scared off.
86 * - By default, in versions before 0.9.10, unused portions of the data
87 * file might receive garbage data from memory freed by other code.
88 * (This does not happen when using the #MDB_WRITEMAP flag.) As of
89 * 0.9.10 the default behavior is to initialize such memory before
90 * writing to the data file. Since there may be a slight performance
91 * cost due to this initialization, applications may disable it using
92 * the #MDB_NOMEMINIT flag. Applications handling sensitive data
93 * which must not be written should not use this flag. This flag is
94 * irrelevant when using #MDB_WRITEMAP.
96 * - A thread can only use one transaction at a time, plus any child
97 * transactions. Each transaction belongs to one thread. See below.
98 * The #MDB_NOTLS flag changes this for read-only transactions.
100 * - Use an MDB_env* in the process which opened it, without fork()ing.
102 * - Do not have open an LMDB database twice in the same process at
103 * the same time. Not even from a plain open() call - close()ing it
104 * breaks flock() advisory locking.
106 * - Avoid long-lived transactions. Read transactions prevent
107 * reuse of pages freed by newer write transactions, thus the
108 * database can grow quickly. Write transactions prevent
109 * other write transactions, since writes are serialized.
111 * - Avoid suspending a process with active transactions. These
112 * would then be "long-lived" as above. Also read transactions
113 * suspended when writers commit could sometimes see wrong data.
115 * ...when several processes can use a database concurrently:
117 * - Avoid aborting a process with an active transaction.
118 * The transaction becomes "long-lived" as above until a check
119 * for stale readers is performed or the lockfile is reset,
120 * since the process may not remove it from the lockfile.
122 * This does not apply to write transactions if the system clears
123 * stale writers, see above.
125 * - If you do that anyway, do a periodic check for stale readers. Or
126 * close the environment once in a while, so the lockfile can get reset.
128 * - Do not use LMDB databases on remote filesystems, even between
129 * processes on the same host. This breaks flock() on some OSes,
130 * possibly memory map sync, and certainly sync between programs
131 * on different hosts.
133 * - Opening a database can fail if another process is opening or
134 * closing it at exactly the same time.
136 * @author Howard Chu, Symas Corporation.
138 * @copyright Copyright 2011-2016 Howard Chu, Symas Corp. All rights reserved.
140 * Redistribution and use in source and binary forms, with or without
141 * modification, are permitted only as authorized by the OpenLDAP
144 * A copy of this license is available in the file LICENSE in the
145 * top-level directory of the distribution or, alternatively, at
146 * <http://www.OpenLDAP.org/license.html>.
149 * This code is derived from btree.c written by Martin Hedenfalk.
151 * Copyright (c) 2009, 2010 Martin Hedenfalk <martin@bzero.se>
153 * Permission to use, copy, modify, and distribute this software for any
154 * purpose with or without fee is hereby granted, provided that the above
155 * copyright notice and this permission notice appear in all copies.
157 * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
158 * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
159 * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
160 * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
161 * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
162 * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
163 * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
168 #include <sys/types.h>
169 #include <inttypes.h>
175 /** Unix permissions for creating files, or dummy definition for Windows */
177 typedef int mdb_mode_t;
179 typedef mode_t mdb_mode_t;
183 # define MDB_FMT_Z "I"
185 # define MDB_FMT_Z "z" /**< printf/scanf format modifier for size_t */
189 typedef uint64_t mdb_size_t;
190 #define MDB_SIZE_MAX UINT64_MAX
192 # define MDB_FMT_Y "I64"
194 # define MDB_FMT_Y "ll"
196 #define mdb_env_create mdb_env_create_vl32 /**< Prevent mixing with non-VL32 builds */
198 typedef size_t mdb_size_t;
199 # define MDB_SIZE_MAX SIZE_MAX /**< max #mdb_size_t */
200 # define MDB_FMT_Y MDB_FMT_Z /**< Obsolescent, see #MDB_PRIz()/#MDB_SCNz() */
203 /** #mdb_size_t printf formats, \b t = one of [diouxX] without quotes */
204 #define MDB_PRIz(t) MDB_FMT_Y #t
205 /** #mdb_size_t scanf formats, \b t = one of [dioux] without quotes */
206 #define MDB_SCNz(t) MDB_FMT_Y #t
208 /** An abstraction for a file handle.
209 * On POSIX systems file handles are small integers. On Windows
210 * they're opaque pointers.
213 typedef void *mdb_filehandle_t;
215 typedef int mdb_filehandle_t;
218 /** @defgroup mdb LMDB API
220 * @brief OpenLDAP Lightning Memory-Mapped Database Manager
222 /** @defgroup Version Version Macros
225 /** Library major version */
226 #define MDB_VERSION_MAJOR 0
227 /** Library minor version */
228 #define MDB_VERSION_MINOR 9
229 /** Library patch version */
230 #define MDB_VERSION_PATCH 70
232 /** Combine args a,b,c into a single integer for easy version comparisons */
233 #define MDB_VERINT(a,b,c) (((a) << 24) | ((b) << 16) | (c))
235 /** The full library version as a single integer */
236 #define MDB_VERSION_FULL \
237 MDB_VERINT(MDB_VERSION_MAJOR,MDB_VERSION_MINOR,MDB_VERSION_PATCH)
239 /** The release date of this library version */
240 #define MDB_VERSION_DATE "December 19, 2015"
242 /** A stringifier for the version info */
243 #define MDB_VERSTR(a,b,c,d) "LMDB " #a "." #b "." #c ": (" d ")"
245 /** A helper for the stringifier macro */
246 #define MDB_VERFOO(a,b,c,d) MDB_VERSTR(a,b,c,d)
248 /** The full library version as a C string */
249 #define MDB_VERSION_STRING \
250 MDB_VERFOO(MDB_VERSION_MAJOR,MDB_VERSION_MINOR,MDB_VERSION_PATCH,MDB_VERSION_DATE)
253 /** @brief Opaque structure for a database environment.
255 * A DB environment supports multiple databases, all residing in the same
258 typedef struct MDB_env MDB_env;
260 /** @brief Opaque structure for a transaction handle.
262 * All database operations require a transaction handle. Transactions may be
263 * read-only or read-write.
265 typedef struct MDB_txn MDB_txn;
267 /** @brief A handle for an individual database in the DB environment. */
268 typedef unsigned int MDB_dbi;
270 /** @brief Opaque structure for navigating through a database */
271 typedef struct MDB_cursor MDB_cursor;
273 /** @brief Generic structure used for passing keys and data in and out
276 * Values returned from the database are valid only until a subsequent
277 * update operation, or the end of the transaction. Do not modify or
278 * free them, they commonly point into the database itself.
280 * Key sizes must be between 1 and #mdb_env_get_maxkeysize() inclusive.
281 * The same applies to data sizes in databases with the #MDB_DUPSORT flag.
282 * Other data items can in theory be from 0 to 0xffffffff bytes long.
284 typedef struct MDB_val {
285 size_t mv_size; /**< size of the data item */
286 void *mv_data; /**< address of the data item */
289 /** @brief A callback function used to compare two keys in a database */
290 typedef int (MDB_cmp_func)(const MDB_val *a, const MDB_val *b);
292 /** @brief A callback function used to relocate a position-dependent data item
293 * in a fixed-address database.
295 * The \b newptr gives the item's desired address in
296 * the memory map, and \b oldptr gives its previous address. The item's actual
297 * data resides at the address in \b item. This callback is expected to walk
298 * through the fields of the record in \b item and modify any
299 * values based at the \b oldptr address to be relative to the \b newptr address.
300 * @param[in,out] item The item that is to be relocated.
301 * @param[in] oldptr The previous address.
302 * @param[in] newptr The new address to relocate to.
303 * @param[in] relctx An application-provided context, set by #mdb_set_relctx().
304 * @todo This feature is currently unimplemented.
306 typedef void (MDB_rel_func)(MDB_val *item, void *oldptr, void *newptr, void *relctx);
308 /** @defgroup mdb_env Environment Flags
311 /** mmap at a fixed address (experimental) */
312 #define MDB_FIXEDMAP 0x01
313 /** no environment directory */
314 #define MDB_NOSUBDIR 0x4000
315 /** don't fsync after commit */
316 #define MDB_NOSYNC 0x10000
318 #define MDB_RDONLY 0x20000
319 /** don't fsync metapage after commit */
320 #define MDB_NOMETASYNC 0x40000
321 /** use writable mmap */
322 #define MDB_WRITEMAP 0x80000
323 /** use asynchronous msync when #MDB_WRITEMAP is used */
324 #define MDB_MAPASYNC 0x100000
325 /** tie reader locktable slots to #MDB_txn objects instead of to threads */
326 #define MDB_NOTLS 0x200000
327 /** don't do any locking, caller must manage their own locks */
328 #define MDB_NOLOCK 0x400000
329 /** don't do readahead (no effect on Windows) */
330 #define MDB_NORDAHEAD 0x800000
331 /** don't initialize malloc'd memory before writing to datafile */
332 #define MDB_NOMEMINIT 0x1000000
335 /** @defgroup mdb_dbi_open Database Flags
338 /** use reverse string keys */
339 #define MDB_REVERSEKEY 0x02
340 /** use sorted duplicates */
341 #define MDB_DUPSORT 0x04
342 /** numeric keys in native byte order: either unsigned int or size_t.
343 * The keys must all be of the same size. */
344 #define MDB_INTEGERKEY 0x08
345 /** with #MDB_DUPSORT, sorted dup items have fixed size */
346 #define MDB_DUPFIXED 0x10
347 /** with #MDB_DUPSORT, dups are #MDB_INTEGERKEY-style integers */
348 #define MDB_INTEGERDUP 0x20
349 /** with #MDB_DUPSORT, use reverse string dups */
350 #define MDB_REVERSEDUP 0x40
351 /** create DB if not already existing */
352 #define MDB_CREATE 0x40000
355 /** @defgroup mdb_put Write Flags
358 /** For put: Don't write if the key already exists. */
359 #define MDB_NOOVERWRITE 0x10
360 /** Only for #MDB_DUPSORT<br>
361 * For put: don't write if the key and data pair already exist.<br>
362 * For mdb_cursor_del: remove all duplicate data items.
364 #define MDB_NODUPDATA 0x20
365 /** For mdb_cursor_put: overwrite the current key/data pair */
366 #define MDB_CURRENT 0x40
367 /** For put: Just reserve space for data, don't copy it. Return a
368 * pointer to the reserved space.
370 #define MDB_RESERVE 0x10000
371 /** Data is being appended, don't split full pages. */
372 #define MDB_APPEND 0x20000
373 /** Duplicate data is being appended, don't split full pages. */
374 #define MDB_APPENDDUP 0x40000
375 /** Store multiple data items in one call. Only for #MDB_DUPFIXED. */
376 #define MDB_MULTIPLE 0x80000
379 /** @defgroup mdb_copy Copy Flags
382 /** Compacting copy: Omit free space from copy, and renumber all
383 * pages sequentially.
385 #define MDB_CP_COMPACT 0x01
388 /** @brief Cursor Get operations.
390 * This is the set of all operations for retrieving data
393 typedef enum MDB_cursor_op {
394 MDB_FIRST, /**< Position at first key/data item */
395 MDB_FIRST_DUP, /**< Position at first data item of current key.
396 Only for #MDB_DUPSORT */
397 MDB_GET_BOTH, /**< Position at key/data pair. Only for #MDB_DUPSORT */
398 MDB_GET_BOTH_RANGE, /**< position at key, nearest data. Only for #MDB_DUPSORT */
399 MDB_GET_CURRENT, /**< Return key/data at current cursor position */
400 MDB_GET_MULTIPLE, /**< Return key and up to a page of duplicate data items
401 from current cursor position. Move cursor to prepare
402 for #MDB_NEXT_MULTIPLE. Only for #MDB_DUPFIXED */
403 MDB_LAST, /**< Position at last key/data item */
404 MDB_LAST_DUP, /**< Position at last data item of current key.
405 Only for #MDB_DUPSORT */
406 MDB_NEXT, /**< Position at next data item */
407 MDB_NEXT_DUP, /**< Position at next data item of current key.
408 Only for #MDB_DUPSORT */
409 MDB_NEXT_MULTIPLE, /**< Return key and up to a page of duplicate data items
410 from next cursor position. Move cursor to prepare
411 for #MDB_NEXT_MULTIPLE. Only for #MDB_DUPFIXED */
412 MDB_NEXT_NODUP, /**< Position at first data item of next key */
413 MDB_PREV, /**< Position at previous data item */
414 MDB_PREV_DUP, /**< Position at previous data item of current key.
415 Only for #MDB_DUPSORT */
416 MDB_PREV_NODUP, /**< Position at last data item of previous key */
417 MDB_SET, /**< Position at specified key */
418 MDB_SET_KEY, /**< Position at specified key, return key + data */
419 MDB_SET_RANGE, /**< Position at first key greater than or equal to specified key. */
420 MDB_PREV_MULTIPLE /**< Position at previous page and return key and up to
421 a page of duplicate data items. Only for #MDB_DUPFIXED */
424 /** @defgroup errors Return Codes
426 * BerkeleyDB uses -30800 to -30999, we'll go under them
429 /** Successful result */
430 #define MDB_SUCCESS 0
431 /** key/data pair already exists */
432 #define MDB_KEYEXIST (-30799)
433 /** key/data pair not found (EOF) */
434 #define MDB_NOTFOUND (-30798)
435 /** Requested page not found - this usually indicates corruption */
436 #define MDB_PAGE_NOTFOUND (-30797)
437 /** Located page was wrong type */
438 #define MDB_CORRUPTED (-30796)
439 /** Update of meta page failed or environment had fatal error */
440 #define MDB_PANIC (-30795)
441 /** Environment version mismatch */
442 #define MDB_VERSION_MISMATCH (-30794)
443 /** File is not a valid LMDB file */
444 #define MDB_INVALID (-30793)
445 /** Environment mapsize reached */
446 #define MDB_MAP_FULL (-30792)
447 /** Environment maxdbs reached */
448 #define MDB_DBS_FULL (-30791)
449 /** Environment maxreaders reached */
450 #define MDB_READERS_FULL (-30790)
451 /** Too many TLS keys in use - Windows only */
452 #define MDB_TLS_FULL (-30789)
453 /** Txn has too many dirty pages */
454 #define MDB_TXN_FULL (-30788)
455 /** Cursor stack too deep - internal error */
456 #define MDB_CURSOR_FULL (-30787)
457 /** Page has not enough space - internal error */
458 #define MDB_PAGE_FULL (-30786)
459 /** Database contents grew beyond environment mapsize */
460 #define MDB_MAP_RESIZED (-30785)
461 /** Operation and DB incompatible, or DB type changed. This can mean:
463 * <li>The operation expects an #MDB_DUPSORT / #MDB_DUPFIXED database.
464 * <li>Opening a named DB when the unnamed DB has #MDB_DUPSORT / #MDB_INTEGERKEY.
465 * <li>Accessing a data record as a database, or vice versa.
466 * <li>The database was dropped and recreated with different flags.
469 #define MDB_INCOMPATIBLE (-30784)
470 /** Invalid reuse of reader locktable slot */
471 #define MDB_BAD_RSLOT (-30783)
472 /** Transaction must abort, has a child, or is invalid */
473 #define MDB_BAD_TXN (-30782)
474 /** Unsupported size of key/DB name/data, or wrong DUPFIXED size */
475 #define MDB_BAD_VALSIZE (-30781)
476 /** The specified DBI was changed unexpectedly */
477 #define MDB_BAD_DBI (-30780)
478 /** Unexpected problem - txn should abort */
479 #define MDB_PROBLEM (-30779)
480 /** The last defined error code */
481 #define MDB_LAST_ERRCODE MDB_PROBLEM
484 /** @brief Statistics for a database in the environment */
485 typedef struct MDB_stat {
486 unsigned int ms_psize; /**< Size of a database page.
487 This is currently the same for all databases. */
488 unsigned int ms_depth; /**< Depth (height) of the B-tree */
489 mdb_size_t ms_branch_pages; /**< Number of internal (non-leaf) pages */
490 mdb_size_t ms_leaf_pages; /**< Number of leaf pages */
491 mdb_size_t ms_overflow_pages; /**< Number of overflow pages */
492 mdb_size_t ms_entries; /**< Number of data items */
495 /** @brief Information about the environment */
496 typedef struct MDB_envinfo {
497 void *me_mapaddr; /**< Address of map, if fixed */
498 mdb_size_t me_mapsize; /**< Size of the data memory map */
499 mdb_size_t me_last_pgno; /**< ID of the last used page */
500 mdb_size_t me_last_txnid; /**< ID of the last committed transaction */
501 unsigned int me_maxreaders; /**< max reader slots in the environment */
502 unsigned int me_numreaders; /**< max reader slots used in the environment */
505 /** @brief Return the LMDB library version information.
507 * @param[out] major if non-NULL, the library major version number is copied here
508 * @param[out] minor if non-NULL, the library minor version number is copied here
509 * @param[out] patch if non-NULL, the library patch version number is copied here
510 * @retval "version string" The library version as a string
512 char *mdb_version(int *major, int *minor, int *patch);
514 /** @brief Return a string describing a given error code.
516 * This function is a superset of the ANSI C X3.159-1989 (ANSI C) strerror(3)
517 * function. If the error code is greater than or equal to 0, then the string
518 * returned by the system function strerror(3) is returned. If the error code
519 * is less than 0, an error string corresponding to the LMDB library error is
520 * returned. See @ref errors for a list of LMDB-specific error codes.
521 * @param[in] err The error code
522 * @retval "error message" The description of the error
524 char *mdb_strerror(int err);
526 /** @brief Create an LMDB environment handle.
528 * This function allocates memory for a #MDB_env structure. To release
529 * the allocated memory and discard the handle, call #mdb_env_close().
530 * Before the handle may be used, it must be opened using #mdb_env_open().
531 * Various other options may also need to be set before opening the handle,
532 * e.g. #mdb_env_set_mapsize(), #mdb_env_set_maxreaders(), #mdb_env_set_maxdbs(),
533 * depending on usage requirements.
534 * @param[out] env The address where the new handle will be stored
535 * @return A non-zero error value on failure and 0 on success.
537 int mdb_env_create(MDB_env **env);
539 /** @brief Open an environment handle.
541 * If this function fails, #mdb_env_close() must be called to discard the #MDB_env handle.
542 * @param[in] env An environment handle returned by #mdb_env_create()
543 * @param[in] path The directory in which the database files reside. This
544 * directory must already exist and be writable.
545 * @param[in] flags Special options for this environment. This parameter
546 * must be set to 0 or by bitwise OR'ing together one or more of the
547 * values described here.
548 * Flags set by mdb_env_set_flags() are also used.
551 * use a fixed address for the mmap region. This flag must be specified
552 * when creating the environment, and is stored persistently in the environment.
553 * If successful, the memory map will always reside at the same virtual address
554 * and pointers used to reference data items in the database will be constant
555 * across multiple invocations. This option may not always work, depending on
556 * how the operating system has allocated memory to shared libraries and other uses.
557 * The feature is highly experimental.
559 * By default, LMDB creates its environment in a directory whose
560 * pathname is given in \b path, and creates its data and lock files
561 * under that directory. With this option, \b path is used as-is for
562 * the database main data file. The database lock file is the \b path
563 * with "-lock" appended.
565 * Open the environment in read-only mode. No write operations will be
566 * allowed. LMDB will still modify the lock file - except on read-only
567 * filesystems, where LMDB does not use locks.
569 * Use a writeable memory map unless MDB_RDONLY is set. This uses
570 * fewer mallocs but loses protection from application bugs
571 * like wild pointer writes and other bad updates into the database.
572 * This may be slightly faster for DBs that fit entirely in RAM, but
573 * is slower for DBs larger than RAM.
574 * Incompatible with nested transactions.
575 * Do not mix processes with and without MDB_WRITEMAP on the same
576 * environment. This can defeat durability (#mdb_env_sync etc).
577 * <li>#MDB_NOMETASYNC
578 * Flush system buffers to disk only once per transaction, omit the
579 * metadata flush. Defer that until the system flushes files to disk,
580 * or next non-MDB_RDONLY commit or #mdb_env_sync(). This optimization
581 * maintains database integrity, but a system crash may undo the last
582 * committed transaction. I.e. it preserves the ACI (atomicity,
583 * consistency, isolation) but not D (durability) database property.
584 * This flag may be changed at any time using #mdb_env_set_flags().
586 * Don't flush system buffers to disk when committing a transaction.
587 * This optimization means a system crash can corrupt the database or
588 * lose the last transactions if buffers are not yet flushed to disk.
589 * The risk is governed by how often the system flushes dirty buffers
590 * to disk and how often #mdb_env_sync() is called. However, if the
591 * filesystem preserves write order and the #MDB_WRITEMAP flag is not
592 * used, transactions exhibit ACI (atomicity, consistency, isolation)
593 * properties and only lose D (durability). I.e. database integrity
594 * is maintained, but a system crash may undo the final transactions.
595 * Note that (#MDB_NOSYNC | #MDB_WRITEMAP) leaves the system with no
596 * hint for when to write transactions to disk, unless #mdb_env_sync()
597 * is called. (#MDB_MAPASYNC | #MDB_WRITEMAP) may be preferable.
598 * This flag may be changed at any time using #mdb_env_set_flags().
600 * When using #MDB_WRITEMAP, use asynchronous flushes to disk.
601 * As with #MDB_NOSYNC, a system crash can then corrupt the
602 * database or lose the last transactions. Calling #mdb_env_sync()
603 * ensures on-disk database integrity until next commit.
604 * This flag may be changed at any time using #mdb_env_set_flags().
606 * Don't use Thread-Local Storage. Tie reader locktable slots to
607 * #MDB_txn objects instead of to threads. I.e. #mdb_txn_reset() keeps
608 * the slot reseved for the #MDB_txn object. A thread may use parallel
609 * read-only transactions. A read-only transaction may span threads if
610 * the user synchronizes its use. Applications that multiplex many
611 * user threads over individual OS threads need this option. Such an
612 * application must also serialize the write transactions in an OS
613 * thread, since LMDB's write locking is unaware of the user threads.
615 * Don't do any locking. If concurrent access is anticipated, the
616 * caller must manage all concurrency itself. For proper operation
617 * the caller must enforce single-writer semantics, and must ensure
618 * that no readers are using old transactions while a writer is
619 * active. The simplest approach is to use an exclusive lock so that
620 * no readers may be active at all when a writer begins.
622 * Turn off readahead. Most operating systems perform readahead on
623 * read requests by default. This option turns it off if the OS
624 * supports it. Turning it off may help random read performance
625 * when the DB is larger than RAM and system RAM is full.
626 * The option is not implemented on Windows.
628 * Don't initialize malloc'd memory before writing to unused spaces
629 * in the data file. By default, memory for pages written to the data
630 * file is obtained using malloc. While these pages may be reused in
631 * subsequent transactions, freshly malloc'd pages will be initialized
632 * to zeroes before use. This avoids persisting leftover data from other
633 * code (that used the heap and subsequently freed the memory) into the
634 * data file. Note that many other system libraries may allocate
635 * and free memory from the heap for arbitrary uses. E.g., stdio may
636 * use the heap for file I/O buffers. This initialization step has a
637 * modest performance cost so some applications may want to disable
638 * it using this flag. This option can be a problem for applications
639 * which handle sensitive data like passwords, and it makes memory
640 * checkers like Valgrind noisy. This flag is not needed with #MDB_WRITEMAP,
641 * which writes directly to the mmap instead of using malloc for pages. The
642 * initialization is also skipped if #MDB_RESERVE is used; the
643 * caller is expected to overwrite all of the memory that was
644 * reserved in that case.
645 * This flag may be changed at any time using #mdb_env_set_flags().
647 * @param[in] mode The UNIX permissions to set on created files and semaphores.
648 * This parameter is ignored on Windows.
649 * @return A non-zero error value on failure and 0 on success. Some possible
652 * <li>#MDB_VERSION_MISMATCH - the version of the LMDB library doesn't match the
653 * version that created the database environment.
654 * <li>#MDB_INVALID - the environment file headers are corrupted.
655 * <li>ENOENT - the directory specified by the path parameter doesn't exist.
656 * <li>EACCES - the user didn't have permission to access the environment files.
657 * <li>EAGAIN - the environment was locked by another process.
660 int mdb_env_open(MDB_env *env, const char *path, unsigned int flags, mdb_mode_t mode);
662 /** @brief Copy an LMDB environment to the specified path.
664 * This function may be used to make a backup of an existing environment.
665 * No lockfile is created, since it gets recreated at need.
666 * @note This call can trigger significant file size growth if run in
667 * parallel with write transactions, because it employs a read-only
668 * transaction. See long-lived transactions under @ref caveats_sec.
669 * @param[in] env An environment handle returned by #mdb_env_create(). It
670 * must have already been opened successfully.
671 * @param[in] path The directory in which the copy will reside. This
672 * directory must already exist and be writable but must otherwise be
674 * @return A non-zero error value on failure and 0 on success.
676 int mdb_env_copy(MDB_env *env, const char *path);
678 /** @brief Copy an LMDB environment to the specified file descriptor.
680 * This function may be used to make a backup of an existing environment.
681 * No lockfile is created, since it gets recreated at need.
682 * @note This call can trigger significant file size growth if run in
683 * parallel with write transactions, because it employs a read-only
684 * transaction. See long-lived transactions under @ref caveats_sec.
685 * @param[in] env An environment handle returned by #mdb_env_create(). It
686 * must have already been opened successfully.
687 * @param[in] fd The filedescriptor to write the copy to. It must
688 * have already been opened for Write access.
689 * @return A non-zero error value on failure and 0 on success.
691 int mdb_env_copyfd(MDB_env *env, mdb_filehandle_t fd);
693 /** @brief Copy an LMDB environment to the specified path, with options.
695 * This function may be used to make a backup of an existing environment.
696 * No lockfile is created, since it gets recreated at need.
697 * @note This call can trigger significant file size growth if run in
698 * parallel with write transactions, because it employs a read-only
699 * transaction. See long-lived transactions under @ref caveats_sec.
700 * @param[in] env An environment handle returned by #mdb_env_create(). It
701 * must have already been opened successfully.
702 * @param[in] path The directory in which the copy will reside. This
703 * directory must already exist and be writable but must otherwise be
705 * @param[in] flags Special options for this operation. This parameter
706 * must be set to 0 or by bitwise OR'ing together one or more of the
707 * values described here.
709 * <li>#MDB_CP_COMPACT - Perform compaction while copying: omit free
710 * pages and sequentially renumber all pages in output. This option
711 * consumes more CPU and runs more slowly than the default.
712 * Currently it fails if the environment has suffered a page leak.
714 * @return A non-zero error value on failure and 0 on success.
716 int mdb_env_copy2(MDB_env *env, const char *path, unsigned int flags);
718 /** @brief Copy an LMDB environment to the specified file descriptor,
721 * This function may be used to make a backup of an existing environment.
722 * No lockfile is created, since it gets recreated at need. See
723 * #mdb_env_copy2() for further details.
724 * @note This call can trigger significant file size growth if run in
725 * parallel with write transactions, because it employs a read-only
726 * transaction. See long-lived transactions under @ref caveats_sec.
727 * @param[in] env An environment handle returned by #mdb_env_create(). It
728 * must have already been opened successfully.
729 * @param[in] fd The filedescriptor to write the copy to. It must
730 * have already been opened for Write access.
731 * @param[in] flags Special options for this operation.
732 * See #mdb_env_copy2() for options.
733 * @return A non-zero error value on failure and 0 on success.
735 int mdb_env_copyfd2(MDB_env *env, mdb_filehandle_t fd, unsigned int flags);
737 /** @brief Return statistics about the LMDB environment.
739 * @param[in] env An environment handle returned by #mdb_env_create()
740 * @param[out] stat The address of an #MDB_stat structure
741 * where the statistics will be copied
743 int mdb_env_stat(MDB_env *env, MDB_stat *stat);
745 /** @brief Return information about the LMDB environment.
747 * @param[in] env An environment handle returned by #mdb_env_create()
748 * @param[out] stat The address of an #MDB_envinfo structure
749 * where the information will be copied
751 int mdb_env_info(MDB_env *env, MDB_envinfo *stat);
753 /** @brief Flush the data buffers to disk.
755 * Data is always written to disk when #mdb_txn_commit() is called,
756 * but the operating system may keep it buffered. LMDB always flushes
757 * the OS buffers upon commit as well, unless the environment was
758 * opened with #MDB_NOSYNC or in part #MDB_NOMETASYNC. This call is
759 * not valid if the environment was opened with #MDB_RDONLY.
760 * @param[in] env An environment handle returned by #mdb_env_create()
761 * @param[in] force If non-zero, force a synchronous flush. Otherwise
762 * if the environment has the #MDB_NOSYNC flag set the flushes
763 * will be omitted, and with #MDB_MAPASYNC they will be asynchronous.
764 * @return A non-zero error value on failure and 0 on success. Some possible
767 * <li>EACCES - the environment is read-only.
768 * <li>EINVAL - an invalid parameter was specified.
769 * <li>EIO - an error occurred during synchronization.
772 int mdb_env_sync(MDB_env *env, int force);
774 /** @brief Close the environment and release the memory map.
776 * Only a single thread may call this function. All transactions, databases,
777 * and cursors must already be closed before calling this function. Attempts to
778 * use any such handles after calling this function will cause a SIGSEGV.
779 * The environment handle will be freed and must not be used again after this call.
780 * @param[in] env An environment handle returned by #mdb_env_create()
782 void mdb_env_close(MDB_env *env);
784 /** @brief Set environment flags.
786 * This may be used to set some flags in addition to those from
787 * #mdb_env_open(), or to unset these flags. If several threads
788 * change the flags at the same time, the result is undefined.
789 * @param[in] env An environment handle returned by #mdb_env_create()
790 * @param[in] flags The flags to change, bitwise OR'ed together
791 * @param[in] onoff A non-zero value sets the flags, zero clears them.
792 * @return A non-zero error value on failure and 0 on success. Some possible
795 * <li>EINVAL - an invalid parameter was specified.
798 int mdb_env_set_flags(MDB_env *env, unsigned int flags, int onoff);
800 /** @brief Get environment flags.
802 * @param[in] env An environment handle returned by #mdb_env_create()
803 * @param[out] flags The address of an integer to store the flags
804 * @return A non-zero error value on failure and 0 on success. Some possible
807 * <li>EINVAL - an invalid parameter was specified.
810 int mdb_env_get_flags(MDB_env *env, unsigned int *flags);
812 /** @brief Return the path that was used in #mdb_env_open().
814 * @param[in] env An environment handle returned by #mdb_env_create()
815 * @param[out] path Address of a string pointer to contain the path. This
816 * is the actual string in the environment, not a copy. It should not be
817 * altered in any way.
818 * @return A non-zero error value on failure and 0 on success. Some possible
821 * <li>EINVAL - an invalid parameter was specified.
824 int mdb_env_get_path(MDB_env *env, const char **path);
826 /** @brief Return the filedescriptor for the given environment.
828 * @param[in] env An environment handle returned by #mdb_env_create()
829 * @param[out] fd Address of a mdb_filehandle_t to contain the descriptor.
830 * @return A non-zero error value on failure and 0 on success. Some possible
833 * <li>EINVAL - an invalid parameter was specified.
836 int mdb_env_get_fd(MDB_env *env, mdb_filehandle_t *fd);
838 /** @brief Set the size of the memory map to use for this environment.
840 * The size should be a multiple of the OS page size. The default is
841 * 10485760 bytes. The size of the memory map is also the maximum size
842 * of the database. The value should be chosen as large as possible,
843 * to accommodate future growth of the database.
844 * This function should be called after #mdb_env_create() and before #mdb_env_open().
845 * It may be called at later times if no transactions are active in
846 * this process. Note that the library does not check for this condition,
847 * the caller must ensure it explicitly.
849 * The new size takes effect immediately for the current process but
850 * will not be persisted to any others until a write transaction has been
851 * committed by the current process. Also, only mapsize increases are
852 * persisted into the environment.
854 * If the mapsize is increased by another process, and data has grown
855 * beyond the range of the current mapsize, #mdb_txn_begin() will
856 * return #MDB_MAP_RESIZED. This function may be called with a size
857 * of zero to adopt the new size.
859 * Any attempt to set a size smaller than the space already consumed
860 * by the environment will be silently changed to the current size of the used space.
861 * @param[in] env An environment handle returned by #mdb_env_create()
862 * @param[in] size The size in bytes
863 * @return A non-zero error value on failure and 0 on success. Some possible
866 * <li>EINVAL - an invalid parameter was specified, or the environment has
867 * an active write transaction.
870 int mdb_env_set_mapsize(MDB_env *env, mdb_size_t size);
872 /** @brief Set the maximum number of threads/reader slots for the environment.
874 * This defines the number of slots in the lock table that is used to track readers in the
875 * the environment. The default is 126.
876 * Starting a read-only transaction normally ties a lock table slot to the
877 * current thread until the environment closes or the thread exits. If
878 * MDB_NOTLS is in use, #mdb_txn_begin() instead ties the slot to the
879 * MDB_txn object until it or the #MDB_env object is destroyed.
880 * This function may only be called after #mdb_env_create() and before #mdb_env_open().
881 * @param[in] env An environment handle returned by #mdb_env_create()
882 * @param[in] readers The maximum number of reader lock table slots
883 * @return A non-zero error value on failure and 0 on success. Some possible
886 * <li>EINVAL - an invalid parameter was specified, or the environment is already open.
889 int mdb_env_set_maxreaders(MDB_env *env, unsigned int readers);
891 /** @brief Get the maximum number of threads/reader slots for the environment.
893 * @param[in] env An environment handle returned by #mdb_env_create()
894 * @param[out] readers Address of an integer to store the number of readers
895 * @return A non-zero error value on failure and 0 on success. Some possible
898 * <li>EINVAL - an invalid parameter was specified.
901 int mdb_env_get_maxreaders(MDB_env *env, unsigned int *readers);
903 /** @brief Set the maximum number of named databases for the environment.
905 * This function is only needed if multiple databases will be used in the
906 * environment. Simpler applications that use the environment as a single
907 * unnamed database can ignore this option.
908 * This function may only be called after #mdb_env_create() and before #mdb_env_open().
910 * Currently a moderate number of slots are cheap but a huge number gets
911 * expensive: 7-120 words per transaction, and every #mdb_dbi_open()
912 * does a linear search of the opened slots.
913 * @param[in] env An environment handle returned by #mdb_env_create()
914 * @param[in] dbs The maximum number of databases
915 * @return A non-zero error value on failure and 0 on success. Some possible
918 * <li>EINVAL - an invalid parameter was specified, or the environment is already open.
921 int mdb_env_set_maxdbs(MDB_env *env, MDB_dbi dbs);
923 /** @brief Get the maximum size of keys and #MDB_DUPSORT data we can write.
925 * Depends on the compile-time constant #MDB_MAXKEYSIZE. Default 511.
927 * @param[in] env An environment handle returned by #mdb_env_create()
928 * @return The maximum size of a key we can write
930 int mdb_env_get_maxkeysize(MDB_env *env);
932 /** @brief Set application information associated with the #MDB_env.
934 * @param[in] env An environment handle returned by #mdb_env_create()
935 * @param[in] ctx An arbitrary pointer for whatever the application needs.
936 * @return A non-zero error value on failure and 0 on success.
938 int mdb_env_set_userctx(MDB_env *env, void *ctx);
940 /** @brief Get the application information associated with the #MDB_env.
942 * @param[in] env An environment handle returned by #mdb_env_create()
943 * @return The pointer set by #mdb_env_set_userctx().
945 void *mdb_env_get_userctx(MDB_env *env);
947 /** @brief A callback function for most LMDB assert() failures,
948 * called before printing the message and aborting.
950 * @param[in] env An environment handle returned by #mdb_env_create().
951 * @param[in] msg The assertion message, not including newline.
953 typedef void MDB_assert_func(MDB_env *env, const char *msg);
955 /** Set or reset the assert() callback of the environment.
956 * Disabled if liblmdb is buillt with NDEBUG.
957 * @note This hack should become obsolete as lmdb's error handling matures.
958 * @param[in] env An environment handle returned by #mdb_env_create().
959 * @param[in] func An #MDB_assert_func function, or 0.
960 * @return A non-zero error value on failure and 0 on success.
962 int mdb_env_set_assert(MDB_env *env, MDB_assert_func *func);
964 /** @brief Create a transaction for use with the environment.
966 * The transaction handle may be discarded using #mdb_txn_abort() or #mdb_txn_commit().
967 * @note A transaction and its cursors must only be used by a single
968 * thread, and a thread may only have a single transaction at a time.
969 * If #MDB_NOTLS is in use, this does not apply to read-only transactions.
970 * @note Cursors may not span transactions.
971 * @param[in] env An environment handle returned by #mdb_env_create()
972 * @param[in] parent If this parameter is non-NULL, the new transaction
973 * will be a nested transaction, with the transaction indicated by \b parent
974 * as its parent. Transactions may be nested to any level. A parent
975 * transaction and its cursors may not issue any other operations than
976 * mdb_txn_commit and mdb_txn_abort while it has active child transactions.
977 * @param[in] flags Special options for this transaction. This parameter
978 * must be set to 0 or by bitwise OR'ing together one or more of the
979 * values described here.
982 * This transaction will not perform any write operations.
984 * Don't flush system buffers to disk when committing this transaction.
985 * <li>#MDB_NOMETASYNC
986 * Flush system buffers but omit metadata flush when committing this transaction.
988 * @param[out] txn Address where the new #MDB_txn handle will be stored
989 * @return A non-zero error value on failure and 0 on success. Some possible
992 * <li>#MDB_PANIC - a fatal error occurred earlier and the environment
994 * <li>#MDB_MAP_RESIZED - another process wrote data beyond this MDB_env's
995 * mapsize and this environment's map must be resized as well.
996 * See #mdb_env_set_mapsize().
997 * <li>#MDB_READERS_FULL - a read-only transaction was requested and
998 * the reader lock table is full. See #mdb_env_set_maxreaders().
999 * <li>ENOMEM - out of memory.
1002 int mdb_txn_begin(MDB_env *env, MDB_txn *parent, unsigned int flags, MDB_txn **txn);
1004 /** @brief Returns the transaction's #MDB_env
1006 * @param[in] txn A transaction handle returned by #mdb_txn_begin()
1008 MDB_env *mdb_txn_env(MDB_txn *txn);
1010 /** @brief Return the transaction's ID.
1012 * This returns the identifier associated with this transaction. For a
1013 * read-only transaction, this corresponds to the snapshot being read;
1014 * concurrent readers will frequently have the same transaction ID.
1016 * @param[in] txn A transaction handle returned by #mdb_txn_begin()
1017 * @return A transaction ID, valid if input is an active transaction.
1019 mdb_size_t mdb_txn_id(MDB_txn *txn);
1021 /** @brief Commit all the operations of a transaction into the database.
1023 * The transaction handle is freed. It and its cursors must not be used
1024 * again after this call, except with #mdb_cursor_renew().
1025 * @note Earlier documentation incorrectly said all cursors would be freed.
1026 * Only write-transactions free cursors.
1027 * @param[in] txn A transaction handle returned by #mdb_txn_begin()
1028 * @return A non-zero error value on failure and 0 on success. Some possible
1031 * <li>EINVAL - an invalid parameter was specified.
1032 * <li>ENOSPC - no more disk space.
1033 * <li>EIO - a low-level I/O error occurred while writing.
1034 * <li>ENOMEM - out of memory.
1037 int mdb_txn_commit(MDB_txn *txn);
1039 /** @brief Abandon all the operations of the transaction instead of saving them.
1041 * The transaction handle is freed. It and its cursors must not be used
1042 * again after this call, except with #mdb_cursor_renew().
1043 * @note Earlier documentation incorrectly said all cursors would be freed.
1044 * Only write-transactions free cursors.
1045 * @param[in] txn A transaction handle returned by #mdb_txn_begin()
1047 void mdb_txn_abort(MDB_txn *txn);
1049 /** @brief Reset a read-only transaction.
1051 * Abort the transaction like #mdb_txn_abort(), but keep the transaction
1052 * handle. #mdb_txn_renew() may reuse the handle. This saves allocation
1053 * overhead if the process will start a new read-only transaction soon,
1054 * and also locking overhead if #MDB_NOTLS is in use. The reader table
1055 * lock is released, but the table slot stays tied to its thread or
1056 * #MDB_txn. Use mdb_txn_abort() to discard a reset handle, and to free
1057 * its lock table slot if MDB_NOTLS is in use.
1058 * Cursors opened within the transaction must not be used
1059 * again after this call, except with #mdb_cursor_renew().
1060 * Reader locks generally don't interfere with writers, but they keep old
1061 * versions of database pages allocated. Thus they prevent the old pages
1062 * from being reused when writers commit new data, and so under heavy load
1063 * the database size may grow much more rapidly than otherwise.
1064 * @param[in] txn A transaction handle returned by #mdb_txn_begin()
1066 void mdb_txn_reset(MDB_txn *txn);
1068 /** @brief Renew a read-only transaction.
1070 * This acquires a new reader lock for a transaction handle that had been
1071 * released by #mdb_txn_reset(). It must be called before a reset transaction
1072 * may be used again.
1073 * @param[in] txn A transaction handle returned by #mdb_txn_begin()
1074 * @return A non-zero error value on failure and 0 on success. Some possible
1077 * <li>#MDB_PANIC - a fatal error occurred earlier and the environment
1078 * must be shut down.
1079 * <li>EINVAL - an invalid parameter was specified.
1082 int mdb_txn_renew(MDB_txn *txn);
1084 /** Compat with version <= 0.9.4, avoid clash with libmdb from MDB Tools project */
1085 #define mdb_open(txn,name,flags,dbi) mdb_dbi_open(txn,name,flags,dbi)
1086 /** Compat with version <= 0.9.4, avoid clash with libmdb from MDB Tools project */
1087 #define mdb_close(env,dbi) mdb_dbi_close(env,dbi)
1089 /** @brief Open a database in the environment.
1091 * A database handle denotes the name and parameters of a database,
1092 * independently of whether such a database exists.
1093 * The database handle may be discarded by calling #mdb_dbi_close().
1094 * The old database handle is returned if the database was already open.
1095 * The handle may only be closed once.
1097 * The database handle will be private to the current transaction until
1098 * the transaction is successfully committed. If the transaction is
1099 * aborted the handle will be closed automatically.
1100 * After a successful commit the handle will reside in the shared
1101 * environment, and may be used by other transactions.
1103 * This function must not be called from multiple concurrent
1104 * transactions in the same process. A transaction that uses
1105 * this function must finish (either commit or abort) before
1106 * any other transaction in the process may use this function.
1108 * To use named databases (with name != NULL), #mdb_env_set_maxdbs()
1109 * must be called before opening the environment. Database names are
1110 * keys in the unnamed database, and may be read but not written.
1112 * @param[in] txn A transaction handle returned by #mdb_txn_begin()
1113 * @param[in] name The name of the database to open. If only a single
1114 * database is needed in the environment, this value may be NULL.
1115 * @param[in] flags Special options for this database. This parameter
1116 * must be set to 0 or by bitwise OR'ing together one or more of the
1117 * values described here.
1119 * <li>#MDB_REVERSEKEY
1120 * Keys are strings to be compared in reverse order, from the end
1121 * of the strings to the beginning. By default, Keys are treated as strings and
1122 * compared from beginning to end.
1124 * Duplicate keys may be used in the database. (Or, from another perspective,
1125 * keys may have multiple data items, stored in sorted order.) By default
1126 * keys must be unique and may have only a single data item.
1127 * <li>#MDB_INTEGERKEY
1128 * Keys are binary integers in native byte order, either unsigned int
1129 * or size_t, and will be sorted as such.
1130 * The keys must all be of the same size.
1132 * This flag may only be used in combination with #MDB_DUPSORT. This option
1133 * tells the library that the data items for this database are all the same
1134 * size, which allows further optimizations in storage and retrieval. When
1135 * all data items are the same size, the #MDB_GET_MULTIPLE and #MDB_NEXT_MULTIPLE
1136 * cursor operations may be used to retrieve multiple items at once.
1137 * <li>#MDB_INTEGERDUP
1138 * This option specifies that duplicate data items are binary integers,
1139 * similar to #MDB_INTEGERKEY keys.
1140 * <li>#MDB_REVERSEDUP
1141 * This option specifies that duplicate data items should be compared as
1142 * strings in reverse order.
1144 * Create the named database if it doesn't exist. This option is not
1145 * allowed in a read-only transaction or a read-only environment.
1147 * @param[out] dbi Address where the new #MDB_dbi handle will be stored
1148 * @return A non-zero error value on failure and 0 on success. Some possible
1151 * <li>#MDB_NOTFOUND - the specified database doesn't exist in the environment
1152 * and #MDB_CREATE was not specified.
1153 * <li>#MDB_DBS_FULL - too many databases have been opened. See #mdb_env_set_maxdbs().
1156 int mdb_dbi_open(MDB_txn *txn, const char *name, unsigned int flags, MDB_dbi *dbi);
1158 /** @brief Retrieve statistics for a database.
1160 * @param[in] txn A transaction handle returned by #mdb_txn_begin()
1161 * @param[in] dbi A database handle returned by #mdb_dbi_open()
1162 * @param[out] stat The address of an #MDB_stat structure
1163 * where the statistics will be copied
1164 * @return A non-zero error value on failure and 0 on success. Some possible
1167 * <li>EINVAL - an invalid parameter was specified.
1170 int mdb_stat(MDB_txn *txn, MDB_dbi dbi, MDB_stat *stat);
1172 /** @brief Retrieve the DB flags for a database handle.
1174 * @param[in] txn A transaction handle returned by #mdb_txn_begin()
1175 * @param[in] dbi A database handle returned by #mdb_dbi_open()
1176 * @param[out] flags Address where the flags will be returned.
1177 * @return A non-zero error value on failure and 0 on success.
1179 int mdb_dbi_flags(MDB_txn *txn, MDB_dbi dbi, unsigned int *flags);
1181 /** @brief Close a database handle. Normally unnecessary. Use with care:
1183 * This call is not mutex protected. Handles should only be closed by
1184 * a single thread, and only if no other threads are going to reference
1185 * the database handle or one of its cursors any further. Do not close
1186 * a handle if an existing transaction has modified its database.
1187 * Doing so can cause misbehavior from database corruption to errors
1188 * like MDB_BAD_VALSIZE (since the DB name is gone).
1190 * Closing a database handle is not necessary, but lets #mdb_dbi_open()
1191 * reuse the handle value. Usually it's better to set a bigger
1192 * #mdb_env_set_maxdbs(), unless that value would be large.
1194 * @param[in] env An environment handle returned by #mdb_env_create()
1195 * @param[in] dbi A database handle returned by #mdb_dbi_open()
1197 void mdb_dbi_close(MDB_env *env, MDB_dbi dbi);
1199 /** @brief Empty or delete+close a database.
1201 * See #mdb_dbi_close() for restrictions about closing the DB handle.
1202 * @param[in] txn A transaction handle returned by #mdb_txn_begin()
1203 * @param[in] dbi A database handle returned by #mdb_dbi_open()
1204 * @param[in] del 0 to empty the DB, 1 to delete it from the
1205 * environment and close the DB handle.
1206 * @return A non-zero error value on failure and 0 on success.
1208 int mdb_drop(MDB_txn *txn, MDB_dbi dbi, int del);
1210 /** @brief Set a custom key comparison function for a database.
1212 * The comparison function is called whenever it is necessary to compare a
1213 * key specified by the application with a key currently stored in the database.
1214 * If no comparison function is specified, and no special key flags were specified
1215 * with #mdb_dbi_open(), the keys are compared lexically, with shorter keys collating
1216 * before longer keys.
1217 * @warning This function must be called before any data access functions are used,
1218 * otherwise data corruption may occur. The same comparison function must be used by every
1219 * program accessing the database, every time the database is used.
1220 * @param[in] txn A transaction handle returned by #mdb_txn_begin()
1221 * @param[in] dbi A database handle returned by #mdb_dbi_open()
1222 * @param[in] cmp A #MDB_cmp_func function
1223 * @return A non-zero error value on failure and 0 on success. Some possible
1226 * <li>EINVAL - an invalid parameter was specified.
1229 int mdb_set_compare(MDB_txn *txn, MDB_dbi dbi, MDB_cmp_func *cmp);
1231 /** @brief Set a custom data comparison function for a #MDB_DUPSORT database.
1233 * This comparison function is called whenever it is necessary to compare a data
1234 * item specified by the application with a data item currently stored in the database.
1235 * This function only takes effect if the database was opened with the #MDB_DUPSORT
1237 * If no comparison function is specified, and no special key flags were specified
1238 * with #mdb_dbi_open(), the data items are compared lexically, with shorter items collating
1239 * before longer items.
1240 * @warning This function must be called before any data access functions are used,
1241 * otherwise data corruption may occur. The same comparison function must be used by every
1242 * program accessing the database, every time the database is used.
1243 * @param[in] txn A transaction handle returned by #mdb_txn_begin()
1244 * @param[in] dbi A database handle returned by #mdb_dbi_open()
1245 * @param[in] cmp A #MDB_cmp_func function
1246 * @return A non-zero error value on failure and 0 on success. Some possible
1249 * <li>EINVAL - an invalid parameter was specified.
1252 int mdb_set_dupsort(MDB_txn *txn, MDB_dbi dbi, MDB_cmp_func *cmp);
1254 /** @brief Set a relocation function for a #MDB_FIXEDMAP database.
1256 * @todo The relocation function is called whenever it is necessary to move the data
1257 * of an item to a different position in the database (e.g. through tree
1258 * balancing operations, shifts as a result of adds or deletes, etc.). It is
1259 * intended to allow address/position-dependent data items to be stored in
1260 * a database in an environment opened with the #MDB_FIXEDMAP option.
1261 * Currently the relocation feature is unimplemented and setting
1262 * this function has no effect.
1263 * @param[in] txn A transaction handle returned by #mdb_txn_begin()
1264 * @param[in] dbi A database handle returned by #mdb_dbi_open()
1265 * @param[in] rel A #MDB_rel_func function
1266 * @return A non-zero error value on failure and 0 on success. Some possible
1269 * <li>EINVAL - an invalid parameter was specified.
1272 int mdb_set_relfunc(MDB_txn *txn, MDB_dbi dbi, MDB_rel_func *rel);
1274 /** @brief Set a context pointer for a #MDB_FIXEDMAP database's relocation function.
1276 * See #mdb_set_relfunc and #MDB_rel_func for more details.
1277 * @param[in] txn A transaction handle returned by #mdb_txn_begin()
1278 * @param[in] dbi A database handle returned by #mdb_dbi_open()
1279 * @param[in] ctx An arbitrary pointer for whatever the application needs.
1280 * It will be passed to the callback function set by #mdb_set_relfunc
1281 * as its \b relctx parameter whenever the callback is invoked.
1282 * @return A non-zero error value on failure and 0 on success. Some possible
1285 * <li>EINVAL - an invalid parameter was specified.
1288 int mdb_set_relctx(MDB_txn *txn, MDB_dbi dbi, void *ctx);
1290 /** @brief Get items from a database.
1292 * This function retrieves key/data pairs from the database. The address
1293 * and length of the data associated with the specified \b key are returned
1294 * in the structure to which \b data refers.
1295 * If the database supports duplicate keys (#MDB_DUPSORT) then the
1296 * first data item for the key will be returned. Retrieval of other
1297 * items requires the use of #mdb_cursor_get().
1299 * @note The memory pointed to by the returned values is owned by the
1300 * database. The caller need not dispose of the memory, and may not
1301 * modify it in any way. For values returned in a read-only transaction
1302 * any modification attempts will cause a SIGSEGV.
1303 * @note Values returned from the database are valid only until a
1304 * subsequent update operation, or the end of the transaction.
1305 * @param[in] txn A transaction handle returned by #mdb_txn_begin()
1306 * @param[in] dbi A database handle returned by #mdb_dbi_open()
1307 * @param[in] key The key to search for in the database
1308 * @param[out] data The data corresponding to the key
1309 * @return A non-zero error value on failure and 0 on success. Some possible
1312 * <li>#MDB_NOTFOUND - the key was not in the database.
1313 * <li>EINVAL - an invalid parameter was specified.
1316 int mdb_get(MDB_txn *txn, MDB_dbi dbi, MDB_val *key, MDB_val *data);
1318 /** @brief Store items into a database.
1320 * This function stores key/data pairs in the database. The default behavior
1321 * is to enter the new key/data pair, replacing any previously existing key
1322 * if duplicates are disallowed, or adding a duplicate data item if
1323 * duplicates are allowed (#MDB_DUPSORT).
1324 * @param[in] txn A transaction handle returned by #mdb_txn_begin()
1325 * @param[in] dbi A database handle returned by #mdb_dbi_open()
1326 * @param[in] key The key to store in the database
1327 * @param[in,out] data The data to store
1328 * @param[in] flags Special options for this operation. This parameter
1329 * must be set to 0 or by bitwise OR'ing together one or more of the
1330 * values described here.
1332 * <li>#MDB_NODUPDATA - enter the new key/data pair only if it does not
1333 * already appear in the database. This flag may only be specified
1334 * if the database was opened with #MDB_DUPSORT. The function will
1335 * return #MDB_KEYEXIST if the key/data pair already appears in the
1337 * <li>#MDB_NOOVERWRITE - enter the new key/data pair only if the key
1338 * does not already appear in the database. The function will return
1339 * #MDB_KEYEXIST if the key already appears in the database, even if
1340 * the database supports duplicates (#MDB_DUPSORT). The \b data
1341 * parameter will be set to point to the existing item.
1342 * <li>#MDB_RESERVE - reserve space for data of the given size, but
1343 * don't copy the given data. Instead, return a pointer to the
1344 * reserved space, which the caller can fill in later - before
1345 * the next update operation or the transaction ends. This saves
1346 * an extra memcpy if the data is being generated later.
1347 * LMDB does nothing else with this memory, the caller is expected
1348 * to modify all of the space requested. This flag must not be
1349 * specified if the database was opened with #MDB_DUPSORT.
1350 * <li>#MDB_APPEND - append the given key/data pair to the end of the
1351 * database. This option allows fast bulk loading when keys are
1352 * already known to be in the correct order. Loading unsorted keys
1353 * with this flag will cause a #MDB_KEYEXIST error.
1354 * <li>#MDB_APPENDDUP - as above, but for sorted dup data.
1356 * @return A non-zero error value on failure and 0 on success. Some possible
1359 * <li>#MDB_MAP_FULL - the database is full, see #mdb_env_set_mapsize().
1360 * <li>#MDB_TXN_FULL - the transaction has too many dirty pages.
1361 * <li>EACCES - an attempt was made to write in a read-only transaction.
1362 * <li>EINVAL - an invalid parameter was specified.
1365 int mdb_put(MDB_txn *txn, MDB_dbi dbi, MDB_val *key, MDB_val *data,
1366 unsigned int flags);
1368 /** @brief Delete items from a database.
1370 * This function removes key/data pairs from the database.
1371 * If the database does not support sorted duplicate data items
1372 * (#MDB_DUPSORT) the data parameter is ignored.
1373 * If the database supports sorted duplicates and the data parameter
1374 * is NULL, all of the duplicate data items for the key will be
1375 * deleted. Otherwise, if the data parameter is non-NULL
1376 * only the matching data item will be deleted.
1377 * This function will return #MDB_NOTFOUND if the specified key/data
1378 * pair is not in the database.
1379 * @param[in] txn A transaction handle returned by #mdb_txn_begin()
1380 * @param[in] dbi A database handle returned by #mdb_dbi_open()
1381 * @param[in] key The key to delete from the database
1382 * @param[in] data The data to delete
1383 * @return A non-zero error value on failure and 0 on success. Some possible
1386 * <li>EACCES - an attempt was made to write in a read-only transaction.
1387 * <li>EINVAL - an invalid parameter was specified.
1390 int mdb_del(MDB_txn *txn, MDB_dbi dbi, MDB_val *key, MDB_val *data);
1392 /** @brief Create a cursor handle.
1394 * A cursor is associated with a specific transaction and database.
1395 * A cursor cannot be used when its database handle is closed. Nor
1396 * when its transaction has ended, except with #mdb_cursor_renew().
1397 * It can be discarded with #mdb_cursor_close().
1398 * A cursor in a write-transaction can be closed before its transaction
1399 * ends, and will otherwise be closed when its transaction ends.
1400 * A cursor in a read-only transaction must be closed explicitly, before
1401 * or after its transaction ends. It can be reused with
1402 * #mdb_cursor_renew() before finally closing it.
1403 * @note Earlier documentation said that cursors in every transaction
1404 * were closed when the transaction committed or aborted.
1405 * @param[in] txn A transaction handle returned by #mdb_txn_begin()
1406 * @param[in] dbi A database handle returned by #mdb_dbi_open()
1407 * @param[out] cursor Address where the new #MDB_cursor handle will be stored
1408 * @return A non-zero error value on failure and 0 on success. Some possible
1411 * <li>EINVAL - an invalid parameter was specified.
1414 int mdb_cursor_open(MDB_txn *txn, MDB_dbi dbi, MDB_cursor **cursor);
1416 /** @brief Close a cursor handle.
1418 * The cursor handle will be freed and must not be used again after this call.
1419 * Its transaction must still be live if it is a write-transaction.
1420 * @param[in] cursor A cursor handle returned by #mdb_cursor_open()
1422 void mdb_cursor_close(MDB_cursor *cursor);
1424 /** @brief Renew a cursor handle.
1426 * A cursor is associated with a specific transaction and database.
1427 * Cursors that are only used in read-only
1428 * transactions may be re-used, to avoid unnecessary malloc/free overhead.
1429 * The cursor may be associated with a new read-only transaction, and
1430 * referencing the same database handle as it was created with.
1431 * This may be done whether the previous transaction is live or dead.
1432 * @param[in] txn A transaction handle returned by #mdb_txn_begin()
1433 * @param[in] cursor A cursor handle returned by #mdb_cursor_open()
1434 * @return A non-zero error value on failure and 0 on success. Some possible
1437 * <li>EINVAL - an invalid parameter was specified.
1440 int mdb_cursor_renew(MDB_txn *txn, MDB_cursor *cursor);
1442 /** @brief Return the cursor's transaction handle.
1444 * @param[in] cursor A cursor handle returned by #mdb_cursor_open()
1446 MDB_txn *mdb_cursor_txn(MDB_cursor *cursor);
1448 /** @brief Return the cursor's database handle.
1450 * @param[in] cursor A cursor handle returned by #mdb_cursor_open()
1452 MDB_dbi mdb_cursor_dbi(MDB_cursor *cursor);
1454 /** @brief Retrieve by cursor.
1456 * This function retrieves key/data pairs from the database. The address and length
1457 * of the key are returned in the object to which \b key refers (except for the
1458 * case of the #MDB_SET option, in which the \b key object is unchanged), and
1459 * the address and length of the data are returned in the object to which \b data
1461 * See #mdb_get() for restrictions on using the output values.
1462 * @param[in] cursor A cursor handle returned by #mdb_cursor_open()
1463 * @param[in,out] key The key for a retrieved item
1464 * @param[in,out] data The data of a retrieved item
1465 * @param[in] op A cursor operation #MDB_cursor_op
1466 * @return A non-zero error value on failure and 0 on success. Some possible
1469 * <li>#MDB_NOTFOUND - no matching key found.
1470 * <li>EINVAL - an invalid parameter was specified.
1473 int mdb_cursor_get(MDB_cursor *cursor, MDB_val *key, MDB_val *data,
1476 /** @brief Store by cursor.
1478 * This function stores key/data pairs into the database.
1479 * The cursor is positioned at the new item, or on failure usually near it.
1480 * @note Earlier documentation incorrectly said errors would leave the
1481 * state of the cursor unchanged.
1482 * @param[in] cursor A cursor handle returned by #mdb_cursor_open()
1483 * @param[in] key The key operated on.
1484 * @param[in] data The data operated on.
1485 * @param[in] flags Options for this operation. This parameter
1486 * must be set to 0 or one of the values described here.
1488 * <li>#MDB_CURRENT - replace the item at the current cursor position.
1489 * The \b key parameter must still be provided, and must match it.
1490 * If using sorted duplicates (#MDB_DUPSORT) the data item must still
1491 * sort into the same place. This is intended to be used when the
1492 * new data is the same size as the old. Otherwise it will simply
1493 * perform a delete of the old record followed by an insert.
1494 * <li>#MDB_NODUPDATA - enter the new key/data pair only if it does not
1495 * already appear in the database. This flag may only be specified
1496 * if the database was opened with #MDB_DUPSORT. The function will
1497 * return #MDB_KEYEXIST if the key/data pair already appears in the
1499 * <li>#MDB_NOOVERWRITE - enter the new key/data pair only if the key
1500 * does not already appear in the database. The function will return
1501 * #MDB_KEYEXIST if the key already appears in the database, even if
1502 * the database supports duplicates (#MDB_DUPSORT).
1503 * <li>#MDB_RESERVE - reserve space for data of the given size, but
1504 * don't copy the given data. Instead, return a pointer to the
1505 * reserved space, which the caller can fill in later - before
1506 * the next update operation or the transaction ends. This saves
1507 * an extra memcpy if the data is being generated later. This flag
1508 * must not be specified if the database was opened with #MDB_DUPSORT.
1509 * <li>#MDB_APPEND - append the given key/data pair to the end of the
1510 * database. No key comparisons are performed. This option allows
1511 * fast bulk loading when keys are already known to be in the
1512 * correct order. Loading unsorted keys with this flag will cause
1513 * a #MDB_KEYEXIST error.
1514 * <li>#MDB_APPENDDUP - as above, but for sorted dup data.
1515 * <li>#MDB_MULTIPLE - store multiple contiguous data elements in a
1516 * single request. This flag may only be specified if the database
1517 * was opened with #MDB_DUPFIXED. The \b data argument must be an
1518 * array of two MDB_vals. The mv_size of the first MDB_val must be
1519 * the size of a single data element. The mv_data of the first MDB_val
1520 * must point to the beginning of the array of contiguous data elements.
1521 * The mv_size of the second MDB_val must be the count of the number
1522 * of data elements to store. On return this field will be set to
1523 * the count of the number of elements actually written. The mv_data
1524 * of the second MDB_val is unused.
1526 * @return A non-zero error value on failure and 0 on success. Some possible
1529 * <li>#MDB_MAP_FULL - the database is full, see #mdb_env_set_mapsize().
1530 * <li>#MDB_TXN_FULL - the transaction has too many dirty pages.
1531 * <li>EACCES - an attempt was made to write in a read-only transaction.
1532 * <li>EINVAL - an invalid parameter was specified.
1535 int mdb_cursor_put(MDB_cursor *cursor, MDB_val *key, MDB_val *data,
1536 unsigned int flags);
1538 /** @brief Delete current key/data pair
1540 * This function deletes the key/data pair to which the cursor refers.
1541 * @param[in] cursor A cursor handle returned by #mdb_cursor_open()
1542 * @param[in] flags Options for this operation. This parameter
1543 * must be set to 0 or one of the values described here.
1545 * <li>#MDB_NODUPDATA - delete all of the data items for the current key.
1546 * This flag may only be specified if the database was opened with #MDB_DUPSORT.
1548 * @return A non-zero error value on failure and 0 on success. Some possible
1551 * <li>EACCES - an attempt was made to write in a read-only transaction.
1552 * <li>EINVAL - an invalid parameter was specified.
1555 int mdb_cursor_del(MDB_cursor *cursor, unsigned int flags);
1557 /** @brief Return count of duplicates for current key.
1559 * This call is only valid on databases that support sorted duplicate
1560 * data items #MDB_DUPSORT.
1561 * @param[in] cursor A cursor handle returned by #mdb_cursor_open()
1562 * @param[out] countp Address where the count will be stored
1563 * @return A non-zero error value on failure and 0 on success. Some possible
1566 * <li>EINVAL - cursor is not initialized, or an invalid parameter was specified.
1569 int mdb_cursor_count(MDB_cursor *cursor, mdb_size_t *countp);
1571 /** @brief Compare two data items according to a particular database.
1573 * This returns a comparison as if the two data items were keys in the
1574 * specified database.
1575 * @param[in] txn A transaction handle returned by #mdb_txn_begin()
1576 * @param[in] dbi A database handle returned by #mdb_dbi_open()
1577 * @param[in] a The first item to compare
1578 * @param[in] b The second item to compare
1579 * @return < 0 if a < b, 0 if a == b, > 0 if a > b
1581 int mdb_cmp(MDB_txn *txn, MDB_dbi dbi, const MDB_val *a, const MDB_val *b);
1583 /** @brief Compare two data items according to a particular database.
1585 * This returns a comparison as if the two items were data items of
1586 * the specified database. The database must have the #MDB_DUPSORT flag.
1587 * @param[in] txn A transaction handle returned by #mdb_txn_begin()
1588 * @param[in] dbi A database handle returned by #mdb_dbi_open()
1589 * @param[in] a The first item to compare
1590 * @param[in] b The second item to compare
1591 * @return < 0 if a < b, 0 if a == b, > 0 if a > b
1593 int mdb_dcmp(MDB_txn *txn, MDB_dbi dbi, const MDB_val *a, const MDB_val *b);
1595 /** @brief A callback function used to print a message from the library.
1597 * @param[in] msg The string to be printed.
1598 * @param[in] ctx An arbitrary context pointer for the callback.
1599 * @return < 0 on failure, >= 0 on success.
1601 typedef int (MDB_msg_func)(const char *msg, void *ctx);
1603 /** @brief Dump the entries in the reader lock table.
1605 * @param[in] env An environment handle returned by #mdb_env_create()
1606 * @param[in] func A #MDB_msg_func function
1607 * @param[in] ctx Anything the message function needs
1608 * @return < 0 on failure, >= 0 on success.
1610 int mdb_reader_list(MDB_env *env, MDB_msg_func *func, void *ctx);
1612 /** @brief Check for stale entries in the reader lock table.
1614 * @param[in] env An environment handle returned by #mdb_env_create()
1615 * @param[out] dead Number of stale slots that were cleared
1616 * @return 0 on success, non-zero on failure.
1618 int mdb_reader_check(MDB_env *env, int *dead);
1624 /** @page tools LMDB Command Line Tools
1625 The following describes the command line tools that are available for LMDB.
1632 #endif /* _LMDB_H_ */