xref: /freebsd/lib/libc/db/man/btree.3 (revision a90b9d0159070121c221b966469c3e36d912bf82)
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.
12.\" 3. Neither the name of the University nor the names of its contributors
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.Dd August 18, 1994
29.Dt BTREE 3
30.Os
31.Sh NAME
32.Nm btree
33.Nd "btree database access method"
34.Sh SYNOPSIS
35.In sys/types.h
36.In db.h
37.Sh DESCRIPTION
38The routine
39.Fn dbopen
40is the library interface to database files.
41One of the supported file formats is
42.Nm
43files.
44The general description of the database access methods is in
45.Xr dbopen 3 ,
46this manual page describes only the
47.Nm
48specific information.
49.Pp
50The
51.Nm
52data structure is a sorted, balanced tree structure storing
53associated key/data pairs.
54.Pp
55The
56.Nm
57access method specific data structure provided to
58.Fn dbopen
59is defined in the
60.In db.h
61include file as follows:
62.Bd -literal
63typedef struct {
64	u_long flags;
65	u_int cachesize;
66	int maxkeypage;
67	int minkeypage;
68	u_int psize;
69	int (*compare)(const DBT *key1, const DBT *key2);
70	size_t (*prefix)(const DBT *key1, const DBT *key2);
71	int lorder;
72} BTREEINFO;
73.Ed
74.Pp
75The elements of this structure are as follows:
76.Bl -tag -width indent
77.It Va flags
78The flag value is specified by
79.Em or Ns 'ing
80any of the following values:
81.Bl -tag -width indent
82.It Dv R_DUP
83Permit duplicate keys in the tree, i.e., permit insertion if the key to be
84inserted already exists in the tree.
85The default behavior, as described in
86.Xr dbopen 3 ,
87is to overwrite a matching key when inserting a new key or to fail if
88the
89.Dv R_NOOVERWRITE
90flag is specified.
91The
92.Dv R_DUP
93flag is overridden by the
94.Dv R_NOOVERWRITE
95flag, and if the
96.Dv R_NOOVERWRITE
97flag is specified, attempts to insert duplicate keys into
98the tree will fail.
99.Pp
100If the database contains duplicate keys, the order of retrieval of
101key/data pairs is undefined if the
102.Va get
103routine is used, however,
104.Va seq
105routine calls with the
106.Dv R_CURSOR
107flag set will always return the logical
108.Dq first
109of any group of duplicate keys.
110.El
111.It Va cachesize
112A suggested maximum size (in bytes) of the memory cache.
113This value is
114.Em only
115advisory, and the access method will allocate more memory rather than fail.
116Since every search examines the root page of the tree, caching the most
117recently used pages substantially improves access time.
118In addition, physical writes are delayed as long as possible, so a moderate
119cache can reduce the number of I/O operations significantly.
120Obviously, using a cache increases (but only increases) the likelihood of
121corruption or lost data if the system crashes while a tree is being modified.
122If
123.Va cachesize
124is 0 (no size is specified) a default cache is used.
125.It Va maxkeypage
126The maximum number of keys which will be stored on any single page.
127Not currently implemented.
128.\" The maximum number of keys which will be stored on any single page.
129.\" Because of the way the
130.\" .Nm
131.\" data structure works,
132.\" .Va maxkeypage
133.\" must always be greater than or equal to 2.
134.\" If
135.\" .Va maxkeypage
136.\" is 0 (no maximum number of keys is specified) the page fill factor is
137.\" made as large as possible (which is almost invariably what is wanted).
138.It Va minkeypage
139The minimum number of keys which will be stored on any single page.
140This value is used to determine which keys will be stored on overflow
141pages, i.e., if a key or data item is longer than the pagesize divided
142by the minkeypage value, it will be stored on overflow pages instead
143of in the page itself.
144If
145.Va minkeypage
146is 0 (no minimum number of keys is specified) a value of 2 is used.
147.It Va psize
148Page size is the size (in bytes) of the pages used for nodes in the tree.
149The minimum page size is 512 bytes and the maximum page size is 64K.
150If
151.Va psize
152is 0 (no page size is specified) a page size is chosen based on the
153underlying file system I/O block size.
154.It Va compare
155Compare is the key comparison function.
156It must return an integer less than, equal to, or greater than zero if the
157first key argument is considered to be respectively less than, equal to,
158or greater than the second key argument.
159The same comparison function must be used on a given tree every time it
160is opened.
161If
162.Va compare
163is
164.Dv NULL
165(no comparison function is specified), the keys are compared
166lexically, with shorter keys considered less than longer keys.
167.It Va prefix
168The
169.Va prefix
170element
171is the prefix comparison function.
172If specified, this routine must return the number of bytes of the second key
173argument which are necessary to determine that it is greater than the first
174key argument.
175If the keys are equal, the key length should be returned.
176Note, the usefulness of this routine is very data dependent, but, in some
177data sets can produce significantly reduced tree sizes and search times.
178If
179.Va prefix
180is
181.Dv NULL
182(no prefix function is specified),
183.Em and
184no comparison function is specified, a default lexical comparison routine
185is used.
186If
187.Va prefix
188is
189.Dv NULL
190and a comparison routine is specified, no prefix comparison is
191done.
192.It Va lorder
193The byte order for integers in the stored database metadata.
194The number should represent the order as an integer; for example,
195big endian order would be the number 4,321.
196If
197.Va lorder
198is 0 (no order is specified) the current host order is used.
199.El
200.Pp
201If the file already exists (and the
202.Dv O_TRUNC
203flag is not specified), the
204values specified for the
205.Va flags , lorder
206and
207.Va psize
208arguments
209are ignored
210in favor of the values used when the tree was created.
211.Pp
212Forward sequential scans of a tree are from the least key to the greatest.
213.Pp
214Space freed up by deleting key/data pairs from the tree is never reclaimed,
215although it is normally made available for reuse.
216This means that the
217.Nm
218storage structure is grow-only.
219The only solutions are to avoid excessive deletions, or to create a fresh
220tree periodically from a scan of an existing one.
221.Pp
222Searches, insertions, and deletions in a
223.Nm
224will all complete in
225O lg base N where base is the average fill factor.
226Often, inserting ordered data into
227.Nm Ns s
228results in a low fill factor.
229This implementation has been modified to make ordered insertion the best
230case, resulting in a much better than normal page fill factor.
231.Sh ERRORS
232The
233.Nm
234access method routines may fail and set
235.Va errno
236for any of the errors specified for the library routine
237.Xr dbopen 3 .
238.Sh SEE ALSO
239.Xr dbopen 3 ,
240.Xr hash 3 ,
241.Xr mpool 3 ,
242.Xr recno 3
243.Rs
244.%T "The Ubiquitous B-tree"
245.%A Douglas Comer
246.%J "ACM Comput. Surv. 11"
247.%N 2
248.%D June 1979
249.%P 121-138
250.Re
251.Rs
252.%A Bayer
253.%A Unterauer
254.%T "Prefix B-trees"
255.%J "ACM Transactions on Database Systems"
256.%N 1
257.%V Vol. 2
258.%D March 1977
259.%P 11-26
260.Re
261.Rs
262.%B "The Art of Computer Programming Vol. 3: Sorting and Searching"
263.%A D. E. Knuth
264.%D 1968
265.%P 471-480
266.Re
267.Sh BUGS
268Only big and little endian byte order is supported.
269