|
1
|
/**
|
|
2
|
* '$RCSfile$'
|
|
3
|
* Purpose: A Class that represents a structured query, and can be
|
|
4
|
* constructed from an XML serialization conforming to
|
|
5
|
* pathquery.dtd. The printSQL() method can be used to print
|
|
6
|
* a SQL serialization of the query.
|
|
7
|
* Copyright: 2000 Regents of the University of California and the
|
|
8
|
* National Center for Ecological Analysis and Synthesis
|
|
9
|
* Authors: Matt Jones
|
|
10
|
*
|
|
11
|
* '$Author: tao $'
|
|
12
|
* '$Date: 2007-04-10 16:31:36 -0700 (Tue, 10 Apr 2007) $'
|
|
13
|
* '$Revision: 3224 $'
|
|
14
|
*
|
|
15
|
* This program is free software; you can redistribute it and/or modify
|
|
16
|
* it under the terms of the GNU General Public License as published by
|
|
17
|
* the Free Software Foundation; either version 2 of the License, or
|
|
18
|
* (at your option) any later version.
|
|
19
|
*
|
|
20
|
* This program is distributed in the hope that it will be useful,
|
|
21
|
* but WITHOUT ANY WARRANTY; without even the implied warranty of
|
|
22
|
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
|
|
23
|
* GNU General Public License for more details.
|
|
24
|
*
|
|
25
|
* You should have received a copy of the GNU General Public License
|
|
26
|
* along with this program; if not, write to the Free Software
|
|
27
|
* Foundation, Inc., 59 Temple Place, Suite 330, Boston, MA 02111-1307 USA
|
|
28
|
*/
|
|
29
|
|
|
30
|
package edu.ucsb.nceas.metacat;
|
|
31
|
|
|
32
|
import java.util.Vector;
|
|
33
|
|
|
34
|
/** a utility class that represents a single term in a query */
|
|
35
|
public class QueryTerm
|
|
36
|
{
|
|
37
|
|
|
38
|
private boolean casesensitive = false;
|
|
39
|
|
|
40
|
private String searchmode = null;
|
|
41
|
|
|
42
|
private String value = null;
|
|
43
|
|
|
44
|
private String pathexpr = null;
|
|
45
|
|
|
46
|
private boolean percentageSymbol = false;
|
|
47
|
|
|
48
|
private int countPercentageSearchItem = 0;
|
|
49
|
|
|
50
|
/**
|
|
51
|
* Construct a new instance of a query term for a free text search (using
|
|
52
|
* the value only)
|
|
53
|
*
|
|
54
|
* @param casesensitive
|
|
55
|
* flag indicating whether case is used to match
|
|
56
|
* @param searchmode
|
|
57
|
* determines what kind of substring match is performed (one of
|
|
58
|
* starts-with|ends-with|contains|matches-exactly)
|
|
59
|
* @param value
|
|
60
|
* the text value to match
|
|
61
|
*/
|
|
62
|
public QueryTerm(boolean casesensitive, String searchmode, String value)
|
|
63
|
{
|
|
64
|
this.casesensitive = casesensitive;
|
|
65
|
this.searchmode = searchmode;
|
|
66
|
this.value = value;
|
|
67
|
}
|
|
68
|
|
|
69
|
/**
|
|
70
|
* Construct a new instance of a query term for a structured search
|
|
71
|
* (matching the value only for those nodes in the pathexpr)
|
|
72
|
*
|
|
73
|
* @param casesensitive
|
|
74
|
* flag indicating whether case is used to match
|
|
75
|
* @param searchmode
|
|
76
|
* determines what kind of substring match is performed (one of
|
|
77
|
* starts-with|ends-with|contains|matches-exactly)
|
|
78
|
* @param value
|
|
79
|
* the text value to match
|
|
80
|
* @param pathexpr
|
|
81
|
* the hierarchical path to the nodes to be searched
|
|
82
|
*/
|
|
83
|
public QueryTerm(boolean casesensitive, String searchmode, String value,
|
|
84
|
String pathexpr)
|
|
85
|
{
|
|
86
|
this(casesensitive, searchmode, value);
|
|
87
|
this.pathexpr = pathexpr;
|
|
88
|
}
|
|
89
|
|
|
90
|
/** determine if the QueryTerm is case sensitive */
|
|
91
|
public boolean isCaseSensitive()
|
|
92
|
{
|
|
93
|
return casesensitive;
|
|
94
|
}
|
|
95
|
|
|
96
|
/** get the searchmode parameter */
|
|
97
|
public String getSearchMode()
|
|
98
|
{
|
|
99
|
return searchmode;
|
|
100
|
}
|
|
101
|
|
|
102
|
/** get the Value parameter */
|
|
103
|
public String getValue()
|
|
104
|
{
|
|
105
|
return value;
|
|
106
|
}
|
|
107
|
|
|
108
|
/** get the path expression parameter */
|
|
109
|
public String getPathExpression()
|
|
110
|
{
|
|
111
|
return pathexpr;
|
|
112
|
}
|
|
113
|
|
|
114
|
/** get the percentage count for one query term */
|
|
115
|
public int getPercentageSymbolCount()
|
|
116
|
{
|
|
117
|
return countPercentageSearchItem;
|
|
118
|
}
|
|
119
|
|
|
120
|
/**
|
|
121
|
* create a SQL serialization of the query that this instance represents
|
|
122
|
*/
|
|
123
|
public String printSQL(boolean useXMLIndex)
|
|
124
|
{
|
|
125
|
StringBuffer self = new StringBuffer();
|
|
126
|
|
|
127
|
// Uppercase the search string if case match is not important
|
|
128
|
String casevalue = null;
|
|
129
|
String nodedataterm = null;
|
|
130
|
boolean notEqual = false;
|
|
131
|
if (casesensitive) {
|
|
132
|
nodedataterm = "nodedata";
|
|
133
|
casevalue = value;
|
|
134
|
} else {
|
|
135
|
nodedataterm = "UPPER(nodedata)";
|
|
136
|
casevalue = value.toUpperCase();
|
|
137
|
}
|
|
138
|
|
|
139
|
// Add appropriate wildcards to search string
|
|
140
|
String searchexpr = null;
|
|
141
|
if (searchmode.equals("starts-with")) {
|
|
142
|
searchexpr = nodedataterm + " LIKE '" + casevalue + "%' ";
|
|
143
|
} else if (searchmode.equals("ends-with")) {
|
|
144
|
searchexpr = nodedataterm + " LIKE '%" + casevalue + "' ";
|
|
145
|
} else if (searchmode.equals("contains")) {
|
|
146
|
if (!casevalue.equals("%")) {
|
|
147
|
searchexpr = nodedataterm + " LIKE '%" + casevalue + "%' ";
|
|
148
|
} else {
|
|
149
|
searchexpr = nodedataterm + " LIKE '" + casevalue + "' ";
|
|
150
|
// find percentage symbol
|
|
151
|
percentageSymbol = true;
|
|
152
|
}
|
|
153
|
} else if (searchmode.equals("not-contains")) {
|
|
154
|
notEqual = true;
|
|
155
|
searchexpr = nodedataterm + " LIKE '%" + casevalue + "%' ";
|
|
156
|
} else if (searchmode.equals("equals")) {
|
|
157
|
searchexpr = nodedataterm + " = '" + casevalue + "' ";
|
|
158
|
} else if (searchmode.equals("isnot-equal")) {
|
|
159
|
notEqual = true;
|
|
160
|
searchexpr = nodedataterm + " = '" + casevalue + "' ";
|
|
161
|
} else {
|
|
162
|
String oper = null;
|
|
163
|
if (searchmode.equals("greater-than")) {
|
|
164
|
oper = ">";
|
|
165
|
nodedataterm = "nodedatanumerical";
|
|
166
|
} else if (searchmode.equals("greater-than-equals")) {
|
|
167
|
oper = ">=";
|
|
168
|
nodedataterm = "nodedatanumerical";
|
|
169
|
} else if (searchmode.equals("less-than")) {
|
|
170
|
oper = "<";
|
|
171
|
nodedataterm = "nodedatanumerical";
|
|
172
|
} else if (searchmode.equals("less-than-equals")) {
|
|
173
|
oper = "<=";
|
|
174
|
nodedataterm = "nodedatanumerical";
|
|
175
|
} else {
|
|
176
|
System.out
|
|
177
|
.println("NOT expected case. NOT recognized operator: "
|
|
178
|
+ searchmode);
|
|
179
|
return null;
|
|
180
|
}
|
|
181
|
try {
|
|
182
|
// it is number; numeric comparison
|
|
183
|
// but we need to make sure there is no string in node data
|
|
184
|
searchexpr = nodedataterm + " " + oper + " "
|
|
185
|
+ new Double(casevalue) + " ";
|
|
186
|
} catch (NumberFormatException nfe) {
|
|
187
|
// these are characters; character comparison
|
|
188
|
searchexpr = nodedataterm + " " + oper + " '" + casevalue
|
|
189
|
+ "' ";
|
|
190
|
}
|
|
191
|
}
|
|
192
|
|
|
193
|
|
|
194
|
// to check xml_path_index can be used
|
|
195
|
boolean usePathIndex = false;
|
|
196
|
|
|
197
|
// if pathexpr has been specified in metacat.properties for indexing
|
|
198
|
if(pathexpr != null){
|
|
199
|
if(MetaCatUtil.pathsForIndexing.contains(pathexpr)){
|
|
200
|
usePathIndex = true;
|
|
201
|
}
|
|
202
|
}
|
|
203
|
|
|
204
|
if(usePathIndex){
|
|
205
|
// using xml_path_index table.....
|
|
206
|
if(notEqual == true){
|
|
207
|
self.append("SELECT DISTINCT docid from xml_path_index WHERE");
|
|
208
|
self.append(" docid NOT IN (Select docid FROM xml_path_index WHERE ");
|
|
209
|
self.append(searchexpr);
|
|
210
|
self.append("AND path LIKE '" + pathexpr + "') ");
|
|
211
|
} else {
|
|
212
|
self.append("SELECT DISTINCT docid FROM xml_path_index WHERE ");
|
|
213
|
self.append(searchexpr);
|
|
214
|
self.append("AND path LIKE '" + pathexpr + "' ");
|
|
215
|
}
|
|
216
|
|
|
217
|
} else {
|
|
218
|
// using xml_nodes and xml_index tables
|
|
219
|
|
|
220
|
if(notEqual == true){
|
|
221
|
self.append("SELECT DISTINCT docid from xml_nodes WHERE");
|
|
222
|
self.append(" docid NOT IN (Select docid FROM xml_nodes WHERE ");
|
|
223
|
} else {
|
|
224
|
self.append("(SELECT DISTINCT docid FROM xml_nodes WHERE ");
|
|
225
|
}
|
|
226
|
self.append(searchexpr);
|
|
227
|
|
|
228
|
if (pathexpr != null) {
|
|
229
|
|
|
230
|
// use XML Index
|
|
231
|
if (useXMLIndex) {
|
|
232
|
if (!hasAttributeInPath(pathexpr)) {
|
|
233
|
// without attributes in path
|
|
234
|
self.append("AND parentnodeid IN ");
|
|
235
|
} else {
|
|
236
|
// has a attribute in path
|
|
237
|
String attributeName = QuerySpecification
|
|
238
|
.getAttributeName(pathexpr);
|
|
239
|
self.append(
|
|
240
|
"AND nodetype LIKE 'ATTRIBUTE' AND nodename LIKE '"
|
|
241
|
+ attributeName + "' ");
|
|
242
|
// and the path expression includes element content other than
|
|
243
|
// just './' or '../'
|
|
244
|
if ( (!pathexpr.startsWith(QuerySpecification.
|
|
245
|
ATTRIBUTESYMBOL)) &&
|
|
246
|
(!pathexpr.startsWith("./" +
|
|
247
|
QuerySpecification.ATTRIBUTESYMBOL)) &&
|
|
248
|
(!pathexpr.startsWith("../" +
|
|
249
|
QuerySpecification.ATTRIBUTESYMBOL))) {
|
|
250
|
|
|
251
|
self.append("AND parentnodeid IN ");
|
|
252
|
pathexpr = QuerySpecification
|
|
253
|
.newPathExpressionWithOutAttribute(pathexpr);
|
|
254
|
}
|
|
255
|
}
|
|
256
|
self.append(
|
|
257
|
"(SELECT nodeid FROM xml_index WHERE path LIKE "
|
|
258
|
+ "'" + pathexpr + "') ");
|
|
259
|
}
|
|
260
|
else {
|
|
261
|
// without using XML Index; using nested statements instead
|
|
262
|
self.append("AND parentnodeid IN ");
|
|
263
|
self.append(useNestedStatements(pathexpr));
|
|
264
|
}
|
|
265
|
}
|
|
266
|
else if ( (value.trim()).equals("%")) {
|
|
267
|
//if pathexpr is null and search value is %, is a
|
|
268
|
// percentageSearchItem
|
|
269
|
// the count number will be increase one
|
|
270
|
countPercentageSearchItem++;
|
|
271
|
|
|
272
|
}
|
|
273
|
self.append(") ");
|
|
274
|
}
|
|
275
|
|
|
276
|
return self.toString();
|
|
277
|
}
|
|
278
|
|
|
279
|
/** A method to judge if a path have attribute */
|
|
280
|
private boolean hasAttributeInPath(String path)
|
|
281
|
{
|
|
282
|
if (path.indexOf(QuerySpecification.ATTRIBUTESYMBOL) != -1) {
|
|
283
|
return true;
|
|
284
|
} else {
|
|
285
|
return false;
|
|
286
|
}
|
|
287
|
}
|
|
288
|
|
|
289
|
/*
|
|
290
|
* Constraint the query with @pathexp without using the XML Index, but
|
|
291
|
* nested SQL statements instead. The query migth be slower.
|
|
292
|
*/
|
|
293
|
public static String useNestedStatements(String pathexpr)
|
|
294
|
{
|
|
295
|
System.out.println("pathexpr: " + pathexpr);
|
|
296
|
StringBuffer nestedStmts = new StringBuffer();
|
|
297
|
Vector nodes = new Vector();
|
|
298
|
String path = pathexpr;
|
|
299
|
int inx = 0;
|
|
300
|
|
|
301
|
do {
|
|
302
|
inx = path.lastIndexOf("/");
|
|
303
|
|
|
304
|
nodes.addElement(path.substring(inx + 1));
|
|
305
|
path = path.substring(0, Math.abs(inx));
|
|
306
|
} while (inx > 0);
|
|
307
|
|
|
308
|
// nested statements
|
|
309
|
int i = 0;
|
|
310
|
for (i = 0; i < nodes.size() - 1; i++) {
|
|
311
|
nestedStmts.append("(SELECT nodeid FROM xml_nodes"
|
|
312
|
+ " WHERE nodename LIKE '" + (String) nodes.elementAt(i)
|
|
313
|
+ "'" + " AND parentnodeid IN ");
|
|
314
|
}
|
|
315
|
// for the last statement: it is without " AND parentnodeid IN "
|
|
316
|
nestedStmts.append("(SELECT nodeid FROM xml_nodes"
|
|
317
|
+ " WHERE nodename LIKE '" + (String) nodes.elementAt(i) + "'");
|
|
318
|
// node.size() number of closing brackets
|
|
319
|
for (i = 0; i < nodes.size(); i++) {
|
|
320
|
nestedStmts.append(")");
|
|
321
|
}
|
|
322
|
|
|
323
|
return nestedStmts.toString();
|
|
324
|
}
|
|
325
|
|
|
326
|
/**
|
|
327
|
* create a String description of the query that this instance represents.
|
|
328
|
* This should become a way to get the XML serialization of the query.
|
|
329
|
*/
|
|
330
|
public String toString()
|
|
331
|
{
|
|
332
|
|
|
333
|
return this.printSQL(true);
|
|
334
|
}
|
|
335
|
|
|
336
|
/**
|
|
337
|
* Compare two query terms to see if they have same search value.
|
|
338
|
* @param term
|
|
339
|
* @return
|
|
340
|
*/
|
|
341
|
public boolean hasSameSearchValue(QueryTerm term)
|
|
342
|
{
|
|
343
|
boolean same = false;
|
|
344
|
if (term != null)
|
|
345
|
{
|
|
346
|
String searchValue = term.getValue();
|
|
347
|
if (searchValue != null && this.value != null)
|
|
348
|
{
|
|
349
|
if (searchValue.equalsIgnoreCase(this.value))
|
|
350
|
{
|
|
351
|
same = true;
|
|
352
|
}
|
|
353
|
}
|
|
354
|
}
|
|
355
|
return same;
|
|
356
|
}
|
|
357
|
}
|