1 package org.apache.lucene.document;
4 * Licensed to the Apache Software Foundation (ASF) under one or more
5 * contributor license agreements. See the NOTICE file distributed with
6 * this work for additional information regarding copyright ownership.
7 * The ASF licenses this file to You under the Apache License, Version 2.0
8 * (the "License"); you may not use this file except in compliance with
9 * the License. You may obtain a copy of the License at
11 * http://www.apache.org/licenses/LICENSE-2.0
13 * Unless required by applicable law or agreed to in writing, software
14 * distributed under the License is distributed on an "AS IS" BASIS,
15 * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
16 * See the License for the specific language governing permissions and
17 * limitations under the License.
20 import java.io.Reader;
22 import org.apache.lucene.analysis.TokenStream;
23 import org.apache.lucene.analysis.NumericTokenStream;
24 import org.apache.lucene.index.FieldInfo.IndexOptions;
25 import org.apache.lucene.util.NumericUtils;
26 import org.apache.lucene.search.NumericRangeQuery; // javadocs
27 import org.apache.lucene.search.NumericRangeFilter; // javadocs
28 import org.apache.lucene.search.SortField; // javadocs
29 import org.apache.lucene.search.FieldCache; // javadocs
32 * <p>This class provides a {@link Field} that enables indexing
33 * of numeric values for efficient range filtering and
34 * sorting. Here's an example usage, adding an int value:
36 * document.add(new NumericField(name).setIntValue(value));
39 * For optimal performance, re-use the
40 * <code>NumericField</code> and {@link Document} instance for more than
44 * NumericField field = new NumericField(name);
45 * Document document = new Document();
46 * document.add(field);
48 * for(all documents) {
50 * field.setIntValue(value)
51 * writer.addDocument(document);
56 * <p>The java native types <code>int</code>, <code>long</code>,
57 * <code>float</code> and <code>double</code> are
58 * directly supported. However, any value that can be
59 * converted into these native types can also be indexed.
60 * For example, date/time values represented by a
61 * {@link java.util.Date} can be translated into a long
62 * value using the {@link java.util.Date#getTime} method. If you
63 * don't need millisecond precision, you can quantize the
64 * value, either by dividing the result of
65 * {@link java.util.Date#getTime} or using the separate getters
66 * (for year, month, etc.) to construct an <code>int</code> or
67 * <code>long</code> value.</p>
69 * <p>To perform range querying or filtering against a
70 * <code>NumericField</code>, use {@link NumericRangeQuery} or {@link
71 * NumericRangeFilter}. To sort according to a
72 * <code>NumericField</code>, use the normal numeric sort types, eg
73 * {@link SortField#INT}. <code>NumericField</code> values
74 * can also be loaded directly from {@link FieldCache}.</p>
76 * <p>By default, a <code>NumericField</code>'s value is not stored but
77 * is indexed for range filtering and sorting. You can use
78 * the {@link #NumericField(String,Field.Store,boolean)}
79 * constructor if you need to change these defaults.</p>
81 * <p>You may add the same field name as a <code>NumericField</code> to
82 * the same document more than once. Range querying and
83 * filtering will be the logical OR of all values; so a range query
84 * will hit all documents that have at least one value in
85 * the range. However sort behavior is not defined. If you need to sort,
86 * you should separately index a single-valued <code>NumericField</code>.</p>
88 * <p>A <code>NumericField</code> will consume somewhat more disk space
89 * in the index than an ordinary single-valued field.
90 * However, for a typical index that includes substantial
91 * textual content per document, this increase will likely
92 * be in the noise. </p>
94 * <p>Within Lucene, each numeric value is indexed as a
95 * <em>trie</em> structure, where each term is logically
96 * assigned to larger and larger pre-defined brackets (which
97 * are simply lower-precision representations of the value).
98 * The step size between each successive bracket is called the
99 * <code>precisionStep</code>, measured in bits. Smaller
100 * <code>precisionStep</code> values result in larger number
101 * of brackets, which consumes more disk space in the index
102 * but may result in faster range search performance. The
103 * default value, 4, was selected for a reasonable tradeoff
104 * of disk space consumption versus performance. You can
105 * use the expert constructor {@link
106 * #NumericField(String,int,Field.Store,boolean)} if you'd
107 * like to change the value. Note that you must also
108 * specify a congruent value when creating {@link
109 * NumericRangeQuery} or {@link NumericRangeFilter}.
110 * For low cardinality fields larger precision steps are good.
111 * If the cardinality is < 100, it is fair
112 * to use {@link Integer#MAX_VALUE}, which produces one
115 * <p>For more information on the internals of numeric trie
116 * indexing, including the <a
117 * href="../search/NumericRangeQuery.html#precisionStepDesc"><code>precisionStep</code></a>
118 * configuration, see {@link NumericRangeQuery}. The format of
119 * indexed values is described in {@link NumericUtils}.
121 * <p>If you only need to sort by numeric value, and never
122 * run range querying/filtering, you can index using a
123 * <code>precisionStep</code> of {@link Integer#MAX_VALUE}.
124 * This will minimize disk space consumed. </p>
126 * <p>More advanced users can instead use {@link
127 * NumericTokenStream} directly, when indexing numbers. This
128 * class is a wrapper around this token stream type for
129 * easier, more intuitive usage.</p>
133 public final class NumericField extends AbstractField {
135 /** Data type of the value in {@link NumericField}.
138 public static enum DataType { INT, LONG, FLOAT, DOUBLE }
140 private transient NumericTokenStream numericTS;
141 private DataType type;
142 private final int precisionStep;
145 * Creates a field for numeric values using the default <code>precisionStep</code>
146 * {@link NumericUtils#PRECISION_STEP_DEFAULT} (4). The instance is not yet initialized with
147 * a numeric value, before indexing a document containing this field,
148 * set a value using the various set<em>???</em>Value() methods.
149 * This constructor creates an indexed, but not stored field.
150 * @param name the field name
152 public NumericField(String name) {
153 this(name, NumericUtils.PRECISION_STEP_DEFAULT, Field.Store.NO, true);
157 * Creates a field for numeric values using the default <code>precisionStep</code>
158 * {@link NumericUtils#PRECISION_STEP_DEFAULT} (4). The instance is not yet initialized with
159 * a numeric value, before indexing a document containing this field,
160 * set a value using the various set<em>???</em>Value() methods.
161 * @param name the field name
162 * @param store if the field should be stored, {@link Document#getFieldable}
163 * then returns {@code NumericField} instances on search results.
164 * @param index if the field should be indexed using {@link NumericTokenStream}
166 public NumericField(String name, Field.Store store, boolean index) {
167 this(name, NumericUtils.PRECISION_STEP_DEFAULT, store, index);
171 * Creates a field for numeric values with the specified
172 * <code>precisionStep</code>. The instance is not yet initialized with
173 * a numeric value, before indexing a document containing this field,
174 * set a value using the various set<em>???</em>Value() methods.
175 * This constructor creates an indexed, but not stored field.
176 * @param name the field name
177 * @param precisionStep the used <a href="../search/NumericRangeQuery.html#precisionStepDesc">precision step</a>
179 public NumericField(String name, int precisionStep) {
180 this(name, precisionStep, Field.Store.NO, true);
184 * Creates a field for numeric values with the specified
185 * <code>precisionStep</code>. The instance is not yet initialized with
186 * a numeric value, before indexing a document containing this field,
187 * set a value using the various set<em>???</em>Value() methods.
188 * @param name the field name
189 * @param precisionStep the used <a href="../search/NumericRangeQuery.html#precisionStepDesc">precision step</a>
190 * @param store if the field should be stored, {@link Document#getFieldable}
191 * then returns {@code NumericField} instances on search results.
192 * @param index if the field should be indexed using {@link NumericTokenStream}
194 public NumericField(String name, int precisionStep, Field.Store store, boolean index) {
195 super(name, store, index ? Field.Index.ANALYZED_NO_NORMS : Field.Index.NO, Field.TermVector.NO);
196 this.precisionStep = precisionStep;
197 setIndexOptions(IndexOptions.DOCS_ONLY);
200 /** Returns a {@link NumericTokenStream} for indexing the numeric value. */
201 public TokenStream tokenStreamValue() {
204 if (numericTS == null) {
205 // lazy init the TokenStream as it is heavy to instantiate (attributes,...),
206 // if not needed (stored field loading)
207 numericTS = new NumericTokenStream(precisionStep);
208 // initialize value in TokenStream
209 if (fieldsData != null) {
211 final Number val = (Number) fieldsData;
214 numericTS.setIntValue(val.intValue()); break;
216 numericTS.setLongValue(val.longValue()); break;
218 numericTS.setFloatValue(val.floatValue()); break;
220 numericTS.setDoubleValue(val.doubleValue()); break;
222 assert false : "Should never get here";
229 /** Returns always <code>null</code> for numeric fields */
231 public byte[] getBinaryValue(byte[] result){
235 /** Returns always <code>null</code> for numeric fields */
236 public Reader readerValue() {
240 /** Returns the numeric value as a string. This format is also returned if you call {@link Document#get(String)}
241 * on search results. It is recommended to use {@link Document#getFieldable} instead
242 * that returns {@code NumericField} instances. You can then use {@link #getNumericValue}
243 * to return the stored value. */
244 public String stringValue() {
245 return (fieldsData == null) ? null : fieldsData.toString();
248 /** Returns the current numeric value as a subclass of {@link Number}, <code>null</code> if not yet initialized. */
249 public Number getNumericValue() {
250 return (Number) fieldsData;
253 /** Returns the precision step. */
254 public int getPrecisionStep() {
255 return precisionStep;
258 /** Returns the data type of the current value, {@code null} if not yet set.
261 public DataType getDataType() {
266 * Initializes the field with the supplied <code>long</code> value.
267 * @param value the numeric value
268 * @return this instance, because of this you can use it the following way:
269 * <code>document.add(new NumericField(name, precisionStep).setLongValue(value))</code>
271 public NumericField setLongValue(final long value) {
272 if (numericTS != null) numericTS.setLongValue(value);
273 fieldsData = Long.valueOf(value);
274 type = DataType.LONG;
279 * Initializes the field with the supplied <code>int</code> value.
280 * @param value the numeric value
281 * @return this instance, because of this you can use it the following way:
282 * <code>document.add(new NumericField(name, precisionStep).setIntValue(value))</code>
284 public NumericField setIntValue(final int value) {
285 if (numericTS != null) numericTS.setIntValue(value);
286 fieldsData = Integer.valueOf(value);
292 * Initializes the field with the supplied <code>double</code> value.
293 * @param value the numeric value
294 * @return this instance, because of this you can use it the following way:
295 * <code>document.add(new NumericField(name, precisionStep).setDoubleValue(value))</code>
297 public NumericField setDoubleValue(final double value) {
298 if (numericTS != null) numericTS.setDoubleValue(value);
299 fieldsData = Double.valueOf(value);
300 type = DataType.DOUBLE;
305 * Initializes the field with the supplied <code>float</code> value.
306 * @param value the numeric value
307 * @return this instance, because of this you can use it the following way:
308 * <code>document.add(new NumericField(name, precisionStep).setFloatValue(value))</code>
310 public NumericField setFloatValue(final float value) {
311 if (numericTS != null) numericTS.setFloatValue(value);
312 fieldsData = Float.valueOf(value);
313 type = DataType.FLOAT;