- Uniformly use .In for header file references.
[dragonfly.git] / lib / libc / db / man / btree.3
CommitLineData
984263bc
MD
1.\" Copyright (c) 1990, 1993
2.\" The Regents of the University of California. All rights reserved.
3.\"
4.\" Redistribution and use in source and binary forms, with or without
5.\" modification, are permitted provided that the following conditions
6.\" are met:
7.\" 1. Redistributions of source code must retain the above copyright
8.\" notice, this list of conditions and the following disclaimer.
9.\" 2. Redistributions in binary form must reproduce the above copyright
10.\" notice, this list of conditions and the following disclaimer in the
11.\" documentation and/or other materials provided with the distribution.
cbebfd39 12.\" 3. Neither the name of the University nor the names of its contributors
984263bc
MD
13.\" may be used to endorse or promote products derived from this software
14.\" without specific prior written permission.
15.\"
16.\" THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
17.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
18.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
19.\" ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
20.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
21.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
22.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
23.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
24.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
25.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
26.\" SUCH DAMAGE.
27.\"
28.\" @(#)btree.3 8.4 (Berkeley) 8/18/94
29.\" $FreeBSD: src/lib/libc/db/man/btree.3,v 1.3.2.3 2003/03/15 15:11:05 trhodes Exp $
44cb301e 30.\" $DragonFly: src/lib/libc/db/man/btree.3,v 1.4 2006/05/26 19:39:36 swildner Exp $
984263bc
MD
31.\"
32.Dd August 18, 1994
33.Dt BTREE 3
34.Os
35.Sh NAME
36.Nm btree
37.Nd "btree database access method"
38.Sh SYNOPSIS
39.In sys/types.h
40.In db.h
41.Sh DESCRIPTION
42The routine
43.Fn dbopen
44is the library interface to database files.
45One of the supported file formats is
46.Nm
47files.
48The general description of the database access methods is in
49.Xr dbopen 3 ,
50this manual page describes only the
51.Nm
52specific information.
53.Pp
54The
55.Nm
56data structure is a sorted, balanced tree structure storing
57associated key/data pairs.
58.Pp
59The
60.Nm
61access method specific data structure provided to
62.Fn dbopen
63is defined in the
44cb301e 64.In db.h
984263bc
MD
65include file as follows:
66.Bd -literal
67typedef struct {
68 u_long flags;
69 u_int cachesize;
70 int maxkeypage;
71 int minkeypage;
72 u_int psize;
73 int (*compare)(const DBT *key1, const DBT *key2);
74 size_t (*prefix)(const DBT *key1, const DBT *key2);
75 int lorder;
76} BTREEINFO;
77.Ed
78.Pp
79The elements of this structure are as follows:
80.Bl -tag -width indent
81.It Va flags
82The flag value is specified by
83.Em or Ns 'ing
84any of the following values:
85.Bl -tag -width indent
86.It Dv R_DUP
87Permit duplicate keys in the tree, i.e. permit insertion if the key to be
88inserted already exists in the tree.
89The default behavior, as described in
90.Xr dbopen 3 ,
91is to overwrite a matching key when inserting a new key or to fail if
92the
93.Dv R_NOOVERWRITE
94flag is specified.
95The
96.Dv R_DUP
97flag is overridden by the
98.Dv R_NOOVERWRITE
99flag, and if the
100.Dv R_NOOVERWRITE
101flag is specified, attempts to insert duplicate keys into
102the tree will fail.
103.Pp
104If the database contains duplicate keys, the order of retrieval of
105key/data pairs is undefined if the
106.Va get
107routine is used, however,
108.Va seq
109routine calls with the
110.Dv R_CURSOR
111flag set will always return the logical
112.Dq first
113of any group of duplicate keys.
114.El
115.It Va cachesize
116A suggested maximum size (in bytes) of the memory cache.
117This value is
118.Em only
119advisory, and the access method will allocate more memory rather than fail.
120Since every search examines the root page of the tree, caching the most
121recently used pages substantially improves access time.
122In addition, physical writes are delayed as long as possible, so a moderate
123cache can reduce the number of I/O operations significantly.
124Obviously, using a cache increases (but only increases) the likelihood of
125corruption or lost data if the system crashes while a tree is being modified.
126If
127.Va cachesize
128is 0 (no size is specified) a default cache is used.
129.It Va maxkeypage
130The maximum number of keys which will be stored on any single page.
131Not currently implemented.
132.\" The maximum number of keys which will be stored on any single page.
133.\" Because of the way the
134.\" .Nm
135.\" data structure works,
136.\" .Va maxkeypage
137.\" must always be greater than or equal to 2.
138.\" If
139.\" .Va maxkeypage
140.\" is 0 (no maximum number of keys is specified) the page fill factor is
141.\" made as large as possible (which is almost invariably what is wanted).
142.It Va minkeypage
143The minimum number of keys which will be stored on any single page.
144This value is used to determine which keys will be stored on overflow
145pages, i.e. if a key or data item is longer than the pagesize divided
146by the minkeypage value, it will be stored on overflow pages instead
147of in the page itself.
148If
149.Va minkeypage
150is 0 (no minimum number of keys is specified) a value of 2 is used.
151.It Va psize
152Page size is the size (in bytes) of the pages used for nodes in the tree.
153The minimum page size is 512 bytes and the maximum page size is 64K.
154If
155.Va psize
156is 0 (no page size is specified) a page size is chosen based on the
157underlying file system I/O block size.
158.It Va compare
159Compare is the key comparison function.
160It must return an integer less than, equal to, or greater than zero if the
161first key argument is considered to be respectively less than, equal to,
162or greater than the second key argument.
163The same comparison function must be used on a given tree every time it
164is opened.
165If
166.Va compare
167is
168.Dv NULL
169(no comparison function is specified), the keys are compared
170lexically, with shorter keys considered less than longer keys.
171.It Va prefix
172The
173.Va prefix
174element
175is the prefix comparison function.
176If specified, this routine must return the number of bytes of the second key
177argument which are necessary to determine that it is greater than the first
178key argument.
179If the keys are equal, the key length should be returned.
180Note, the usefulness of this routine is very data dependent, but, in some
181data sets can produce significantly reduced tree sizes and search times.
182If
183.Va prefix
184is
185.Dv NULL
186(no prefix function is specified),
187.Em and
188no comparison function is specified, a default lexical comparison routine
189is used.
190If
191.Va prefix
192is
193.Dv NULL
194and a comparison routine is specified, no prefix comparison is
195done.
196.It Va lorder
197The byte order for integers in the stored database metadata.
198The number should represent the order as an integer; for example,
199big endian order would be the number 4,321.
200If
201.Va lorder
202is 0 (no order is specified) the current host order is used.
203.El
204.Pp
205If the file already exists (and the
206.Dv O_TRUNC
207flag is not specified), the
208values specified for the
209.Va flags , lorder
210and
211.Va psize
212arguments
213are ignored
214in favor of the values used when the tree was created.
215.Pp
216Forward sequential scans of a tree are from the least key to the greatest.
217.Pp
218Space freed up by deleting key/data pairs from the tree is never reclaimed,
219although it is normally made available for reuse.
220This means that the
221.Nm
222storage structure is grow-only.
223The only solutions are to avoid excessive deletions, or to create a fresh
224tree periodically from a scan of an existing one.
225.Pp
226Searches, insertions, and deletions in a
227.Nm
228will all complete in
229O lg base N where base is the average fill factor.
230Often, inserting ordered data into
231.Nm Ns s
232results in a low fill factor.
233This implementation has been modified to make ordered insertion the best
234case, resulting in a much better than normal page fill factor.
235.Sh ERRORS
236The
237.Nm
238access method routines may fail and set
239.Va errno
240for any of the errors specified for the library routine
241.Xr dbopen 3 .
242.Sh SEE ALSO
243.Xr dbopen 3 ,
244.Xr hash 3 ,
245.Xr mpool 3 ,
246.Xr recno 3
247.Rs
248.%T "The Ubiquitous B-tree"
249.%A Douglas Comer
250.%J "ACM Comput. Surv. 11"
251.%N 2
252.%D June 1979
253.%P 121-138
254.Re
255.Rs
256.%A Bayer
257.%A Unterauer
258.%T "Prefix B-trees"
259.%J "ACM Transactions on Database Systems"
260.%N 1
261.%V Vol. 2
262.%D March 1977
263.%P 11-26
264.Re
265.Rs
266.%B "The Art of Computer Programming Vol. 3: Sorting and Searching"
267.%A D. E. Knuth
268.%D 1968
269.%P 471-480
270.Re
271.Sh BUGS
272Only big and little endian byte order is supported.