1.\" $NetBSD: btree.3,v 1.13 2012/10/13 15:28:33 njoly Exp $ 2.\" 3.\" Copyright (c) 1990, 1993 4.\" The Regents of the University of California. All rights reserved. 5.\" 6.\" Redistribution and use in source and binary forms, with or without 7.\" modification, are permitted provided that the following conditions 8.\" are met: 9.\" 1. Redistributions of source code must retain the above copyright 10.\" notice, this list of conditions and the following disclaimer. 11.\" 2. Redistributions in binary form must reproduce the above copyright 12.\" notice, this list of conditions and the following disclaimer in the 13.\" documentation and/or other materials provided with the distribution. 14.\" 3. Neither the name of the University nor the names of its contributors 15.\" may be used to endorse or promote products derived from this software 16.\" without specific prior written permission. 17.\" 18.\" THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 19.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 20.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 21.\" ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 22.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 23.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 24.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 25.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 26.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 27.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 28.\" SUCH DAMAGE. 29.\" 30.\" @(#)btree.3 8.4 (Berkeley) 8/18/94 31.\" 32.Dd April 17, 2003 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 btree files. 46The general description of the database access methods is in 47.Xr dbopen 3 , 48this manual page describes only the btree specific information. 49.Pp 50The btree data structure is a sorted, balanced tree structure storing 51associated key/data pairs. 52.Pp 53The btree access method specific data structure provided to 54.Fn dbopen 55is defined in the 56.In db.h 57include file as follows: 58.Bd -literal 59typedef struct { 60 u_long flags; 61 u_int cachesize; 62 int maxkeypage; 63 int minkeypage; 64 u_int psize; 65 int (*compare)(const DBT *key1, const DBT *key2); 66 size_t (*prefix)(const DBT *key1, const DBT *key2); 67 int lorder; 68} BTREEINFO; 69.Ed 70.Pp 71The elements of this structure are as follows: 72.Bl -tag -width maxkeypagex 73.It Fa flags 74The flag value is specified by or'ing any of the following values: 75.Bl -tag -width R_DUP -offset indent 76.It Dv R_DUP 77Permit duplicate keys in the tree, i.e. permit insertion if the key to 78be inserted already exists in the tree. 79The default behavior, as described in 80.Xr dbopen 3 , 81is to overwrite a matching key when inserting a new key or to fail if 82the 83.Dv R_NOOVERWRITE 84flag is specified. 85The 86.Dv R_DUP 87flag is overridden by the 88.Dv R_NOOVERWRITE 89flag, and if the 90.Dv R_NOOVERWRITE 91flag is specified, attempts to insert duplicate keys into the tree 92will fail. 93.Pp 94If the database contains duplicate keys, the order of retrieval of 95key/data pairs is undefined if the 96.Em get 97routine is used, however, 98.Em seq 99routine calls with the 100.Dv R_CURSOR 101flag set will always return the logical 102.Dq first 103of any group of duplicate keys. 104.El 105.It Fa cachesize 106A suggested maximum size (in bytes) of the memory cache. 107This value is 108.Em only 109advisory, and the access method will allocate more memory rather than 110fail. 111Since every search examines the root page of the tree, caching the 112most recently used pages substantially improves access time. 113In addition, physical writes are delayed as long as possible, so a 114moderate cache can reduce the number of I/O operations significantly. 115Obviously, using a cache increases (but only increases) the likelihood 116of corruption or lost data if the system crashes while a tree is being 117modified. 118If 119.Fa cachesize 120is 0 (no size is specified) a default cache is used. 121.It Fa maxkeypage 122The maximum number of keys which will be stored on any single page. 123Not currently implemented. 124.\" The maximum number of keys which will be stored on any single page. 125.\" Because of the way the btree data structure works, 126.\" .Fa maxkeypage 127.\" must always be greater than or equal to 2. 128.\" If 129.\" .Fa maxkeypage 130.\" is 0 (no maximum number of keys is specified) the page fill factor is 131.\" made as large as possible (which is almost invariably what is wanted). 132.It Fa minkeypage 133The minimum number of keys which will be stored on any single page. 134This value is used to determine which keys will be stored on overflow 135pages, i.e., if a key or data item is longer than the pagesize divided 136by the 137.Fa minkeypage 138value, it will be stored on overflow pages instead of in the page 139itself. 140If 141.Fa minkeypage 142is 0 (no minimum number of keys is specified) a value of 2 is used. 143.It Fa psize 144Page size is the size (in bytes) of the pages used for nodes in the 145tree. 146The minimum page size is 512 bytes and the maximum page size is 64K. 147If 148.Fa psize 149is 0 (no page size is specified) a page size is chosen based on the 150underlying file system I/O block size. 151.It Fa compare 152Compare is the key comparison function. 153It must return an integer less than, equal to, or greater than zero if 154the first key argument is considered to be respectively less than, 155equal to, or greater than the second key argument. 156The same comparison function must be used on a given tree every time 157it is opened. 158If 159.Fa compare 160is 161.Dv NULL 162(no comparison function is specified), the keys are compared 163lexically, with shorter keys considered less than longer keys. 164.It Fa prefix 165Prefix is the prefix comparison function. 166If specified, this routine must return the number of bytes of the 167second key argument which are necessary to determine that it is 168greater than the first key argument. 169If the keys are equal, the key length should be returned. 170Note, the usefulness of this routine is very data dependent, but, in 171some data sets can produce significantly reduced tree sizes and search 172times. 173If 174.Fa prefix 175is 176.Dv NULL 177(no prefix function is specified), 178.Em and 179no comparison function is specified, a default lexical comparison 180routine is used. 181If 182.Fa prefix 183is 184.Dv NULL 185and a comparison routine is specified, no prefix comparison is done. 186.It Fa lorder 187The byte order for integers in the stored database metadata. 188The number should represent the order as an integer; for example, 189big endian order would be the number 4,321. 190If 191.Fa lorder 192is 0 (no order is specified) the current host order is used. 193.El 194.Pp 195If the file already exists (and the 196.Dv O_TRUNC 197flag is not specified), the values specified for the parameters flags, 198lorder and psize are ignored in favor of the values used when the tree 199was created. 200.Pp 201Forward sequential scans of a tree are from the least key to the 202greatest. 203.Pp 204Space freed up by deleting key/data pairs from the tree is never 205reclaimed, although it is normally made available for reuse. 206This means that the btree storage structure is grow-only. 207The only solutions are to avoid excessive deletions, or to create a 208fresh tree periodically from a scan of an existing one. 209.Pp 210Searches, insertions, and deletions in a btree will all complete in 211O lg base N where base is the average fill factor. 212Often, inserting ordered data into btrees results in a low fill 213factor. 214This implementation has been modified to make ordered insertion the 215best case, resulting in a much better than normal page fill factor. 216.Sh ERRORS 217The 218.Nm 219access method routines may fail and set 220.Va errno 221for any of the errors specified for the library routine 222.Xr dbopen 3 . 223.Sh SEE ALSO 224.Xr dbopen 3 , 225.Xr hash 3 , 226.Xr mpool 3 , 227.Xr recno 3 228.Rs 229.%T "The Ubiquitous B-tree" 230.%A "Douglas Comer" 231.%J "ACM Comput. Surv." 232.%V 2 233.%N 11 234.%D June 1979 235.%P 121-138 236.Re 237.Rs 238.%T "Prefix B-trees" 239.%A "Bayer" 240.%A "Unterauer" 241.%J "ACM Transactions on Database Systems" 242.%V Vol. 2 243.%N 1 244.%D March 1977 245.%P 11-26 246.Re 247.Rs 248.%B "The Art of Computer Programming Vol. 3: Sorting and Searching" 249.%A "D.E. Knuth" 250.%D 1968 251.%P 471-480 252.Re 253.Sh BUGS 254Only big and little endian byte order is supported. 255