1 <chapter id="fields-and-charsets">
2 <!-- $Id: field-structure.xml,v 1.8 2006-11-28 13:05:57 marc Exp $ -->
3 <title>Field Structure and Character Sets
7 In order to provide a flexible approach to national character set
8 handling, Zebra allows the administrator to configure the set up the
9 system to handle any 8-bit character set — including sets that
10 require multi-octet diacritics or other multi-octet characters. The
11 definition of a character set includes a specification of the
12 permissible values, their sort order (this affects the display in the
13 SCAN function), and relationships between upper- and lowercase
14 characters. Finally, the definition includes the specification of
15 space characters for the set.
19 The operator can define different character sets for different fields,
20 typical examples being standard text fields, numerical fields, and
21 special-purpose fields such as WWW-style linkages (URx).
24 <section id="default-idx-file">
25 <title>The default.idx file</title>
27 The field types, and hence character sets, are associated with data
28 elements by the .abs files (see above).
29 The file <literal>default.idx</literal>
30 provides the association between field type codes (as used in the .abs
31 files) and the character map files (with the .chr suffix). The format
32 of the .idx file is as follows
39 <term>index <replaceable>field type code</replaceable></term>
42 This directive introduces a new search index code.
43 The argument is a one-character code to be used in the
44 .abs files to select this particular index type. An index, roughly,
45 corresponds to a particular structure attribute during search. Refer
46 to <xref linkend="zebrasrv-search"/>.
48 </listitem></varlistentry>
50 <term>sort <replaceable>field code type</replaceable></term>
53 This directive introduces a
54 sort index. The argument is a one-character code to be used in the
55 .abs fie to select this particular index type. The corresponding
56 use attribute must be used in the sort request to refer to this
57 particular sort index. The corresponding character map (see below)
58 is used in the sort process.
60 </listitem></varlistentry>
62 <term>completeness <replaceable>boolean</replaceable></term>
65 This directive enables or disables complete field indexing.
66 The value of the <replaceable>boolean</replaceable> should be 0
67 (disable) or 1. If completeness is enabled, the index entry will
68 contain the complete contents of the field (up to a limit), with words
69 (non-space characters) separated by single space characters
70 (normalized to " " on display). When completeness is
71 disabled, each word is indexed as a separate entry. Complete subfield
72 indexing is most useful for fields which are typically browsed (eg.
73 titles, authors, or subjects), or instances where a match on a
74 complete subfield is essential (eg. exact title searching). For fields
75 where completeness is disabled, the search engine will interpret a
76 search containing space characters as a word proximity search.
78 </listitem></varlistentry>
80 <varlistentry id="default.idx.firstinfield">
81 <term>firstinfield <replaceable>boolean</replaceable></term>
84 This directive enables or disables first-in-field indexing.
85 The value of the <replaceable>boolean</replaceable> should be 0
88 </listitem></varlistentry>
90 <varlistentry id="default.idx.alwaysmatches">
91 <term>alwaysmatches <replaceable>boolean</replaceable></term>
94 This directive enables or disables alwaysmatches indexing.
95 The value of the <replaceable>boolean</replaceable> should be 0
98 </listitem></varlistentry>
101 <term>charmap <replaceable>filename</replaceable></term>
104 This is the filename of the character
105 map to be used for this index for field type.
106 See <xref linkend="character-map-files"/> for details.
108 </listitem></varlistentry>
113 <section id="character-map-files">
114 <title>The character map file format</title>
116 The character map files are used to define the word tokenization
117 and character normalization performed before inserting text into
118 the inverse indexes. Zebra ships with the predefined character map
119 files <filename>tab/*.chr</filename>. Users are allowed to add
120 and/or modify maps according to their needs.
123 <table id="querymodel-attribute-sets-table" frame="top">
124 <title>Character maps predefined in Zebra</title>
128 <entry>File name</entry>
129 <entry>Intended type</entry>
130 <entry>Description</entry>
135 <entry><literal>numeric.chr</literal></entry>
136 <entry><literal>:n</literal></entry>
137 <entry>Numeric digit tokenization and normalization map. All
138 characters not in the set <literal>-{0-9}.,</literal> will be
139 suppressed. Note that floating point numbers are processed
140 fine, but scientific exponential numbers are trashed.</entry>
143 <entry><literal>scan.chr</literal></entry>
144 <entry><literal>:w or :p</literal></entry>
145 <entry>Word tokenization char map for Scandinavian
146 languages. This one resembles the generic word tokenization
147 character map <literal>tab/string.chr</literal>, the main
148 differences are sorting of the special characters
149 <literal>üzæäøöå</literal> and equivalence maps according to
150 Scandinavian language rules.</entry>
153 <entry><literal>string.chr</literal></entry>
154 <entry><literal>:w or :p</literal></entry>
155 <entry>General word tokenization and normalization character
156 map, mostly useful for English texts. Use this to derive your
157 own language tokenization and normalization derivatives.</entry>
160 <entry><literal>urx.chr</literal></entry>
161 <entry><literal>:u</literal></entry>
162 <entry>URL parsing and tokenization character map.</entry>
165 <entry><literal>@</literal></entry>
166 <entry><literal>:0</literal></entry>
167 <entry>Do-nothing character map used for literal binary
168 indexing. There is no existing file associated to it, and
169 there is no normalization or tokenization performed at all.</entry>
176 The contents of the character map files are structured as follows:
180 <term>lowercase <replaceable>value-set</replaceable></term>
183 This directive introduces the basic value set of the field type.
184 The format is an ordered list (without spaces) of the
185 characters which may occur in "words" of the given type.
186 The order of the entries in the list determines the
187 sort order of the index. In addition to single characters, the
188 following combinations are legal:
196 Backslashes may be used to introduce three-digit octal, or
197 two-digit hex representations of single characters
198 (preceded by <literal>x</literal>).
199 In addition, the combinations
200 \\, \\r, \\n, \\t, \\s (space — remember that real
201 space-characters may not occur in the value definition), and
202 \\ are recognized, with their usual interpretation.
208 Curly braces {} may be used to enclose ranges of single
209 characters (possibly using the escape convention described in the
210 preceding point), eg. {a-z} to introduce the
211 standard range of ASCII characters.
212 Note that the interpretation of such a range depends on
213 the concrete representation in your local, physical character set.
219 paranthesises () may be used to enclose multi-byte characters -
220 eg. diacritics or special national combinations (eg. Spanish
221 "ll"). When found in the input stream (or a search term),
222 these characters are viewed and sorted as a single character, with a
223 sorting value depending on the position of the group in the value
232 For example, <literal>scan.chr</literal> contains the following
233 lowercase normalization and sorting order:
235 lowercase {0-9}{a-y}üzæäøöå
238 </listitem></varlistentry>
240 <term>uppercase <replaceable>value-set</replaceable></term>
243 This directive introduces the
244 upper-case equivalences to the value set (if any). The number and
245 order of the entries in the list should be the same as in the
246 <literal>lowercase</literal> directive.
249 For example, <literal>scan.chr</literal> contains the following
250 uppercase equivalent:
252 uppercase {0-9}{A-Y}ÜZÆÄØÖÅ
255 </listitem></varlistentry>
257 <term>space <replaceable>value-set</replaceable></term>
260 This directive introduces the character
261 which separate words in the input stream. Depending on the
262 completeness mode of the field in question, these characters either
263 terminate an index entry, or delimit individual "words" in
264 the input stream. The order of the elements is not significant —
265 otherwise the representation is the same as for the
266 <literal>uppercase</literal> and <literal>lowercase</literal>
270 For example, <literal>scan.chr</literal> contains the following
273 space {\001-\040}!"#$%&'\()*+,-./:;<=>?@\[\\]^_`\{|}~
276 </listitem></varlistentry>
278 <term>map <replaceable>value-set</replaceable>
279 <replaceable>target</replaceable></term>
282 This directive introduces a mapping between each of the
283 members of the value-set on the left to the character on the
284 right. The character on the right must occur in the value
285 set (the <literal>lowercase</literal> directive) of the
286 character set, but it may be a parenthesis-enclosed
287 multi-octet character. This directive may be used to map
288 diacritics to their base characters, or to map HTML-style
289 character-representations to their natural form, etc. The
290 map directive can also be used to ignore leading articles in
291 searching and/or sorting, and to perform other special
292 transformations. See section <xref
293 linkend="leading-articles"/>.
296 For example, <literal>scan.chr</literal> contains the following
297 map instructions among others, to make sure that HTML entity
298 encoded Danish special characters are mapped to the
299 equivalent Latin-1 characters:
306 </listitem></varlistentry>
308 <term>equivalent <replaceable>value-set</replaceable></term>
311 This directive introduces equivalence classes of characters
312 and/or strings for sorting purposes only. It resembles the map
313 directive, but does not affect search and retrieval indexing,
314 but only sorting order under present requests.
317 For example, <literal>scan.chr</literal> contains the following
318 equivalent sorting instructions, which can be uncommented:
326 </listitem></varlistentry>
330 <section id="leading-articles">
331 <title>Ignoring leading articles</title>
333 In addition to specifying sort orders, space (blank) handling,
334 and upper/lowercase folding, you can also use the character map
335 files to make Zebra ignore leading articles in sorting records,
336 or when doing complete field searching.
339 This is done using the <literal>map</literal> directive in the
340 character map file. In a nutshell, what you do is map certain
341 sequences of characters, when they occur <emphasis> in the
342 beginning of a field</emphasis>, to a space. Assuming that the
343 character "@" is defined as a space character in your file, you
349 The effect of these directives is to map either 'the' or 'The',
350 followed by a space character, to a space. The hat ^ character
351 denotes beginning-of-field only when complete-subfield indexing
352 or sort indexing is taking place; otherwise, it is treated just
353 as any other character.
356 Because the <literal>default.idx</literal> file can be used to
357 associate different character maps with different indexing types
358 -- and you can create additional indexing types, should the need
359 arise -- it is possible to specify that leading articles should
360 be ignored either in sorting, in complete-field searching, or
364 If you ignore certain prefixes in sorting, then these will be
365 eliminated from the index, and sorting will take place as if
366 they weren't there. However, if you set the system up to ignore
367 certain prefixes in <emphasis>searching</emphasis>, then these
368 are deleted both from the indexes and from query terms, when the
369 client specifies complete-field searching. This has the effect
370 that a search for 'the science journal' and 'science journal'
371 would both produce the same results.
376 <!-- Keep this comment at the end of the file
381 sgml-minimize-attributes:nil
382 sgml-always-quote-attributes:t
385 sgml-parent-document: "zebra.xml"
386 sgml-local-catalogs: nil
387 sgml-namecase-general:t