公司开发使用的搜索引擎核心是Solr,但是应为业务原因,需要相对复杂权限机制。
1)通过Solr的filterQuery可以实现field过滤,实现过滤项的效果。索引
A{filter1:a,field2:a,field3:a,field4:1}
B{filter1:b,field2:b,field3:b,field4:2}
C{filter1:c,field2:c,field3:c,field4:3}
过滤条件
fq=field:b
结果
B{filter1:b,field2:b,field3:b,field4:b}
2)通过Solr的数字区间,也能方便实现密级的权限方式。索引
A{filter1:a,field2:a,field3:a,field4:1}
B{filter1:b,field2:b,field3:b,field4:2}
C{filter1:c,field2:c,field3:c,field4:3}
过滤条件
fq=field4:[2 TO *]
结果
B{filter1:b,field2:b,field3:b,field4:2}
C{filter1:c,field2:c,field3:c,field4:3}
3)但是真正常用的是多条件的、并集式的过滤项的效果(如:过滤出包涵filed1:a和filed1:b的所有索引)。
在网上搜索了很久,也问了一些做搜索很长时间的同事,都没有找到解决方法,郁闷啊!最后决定自己仔细的把Solr源码读一读,看看有没有解决方法(一直觉得开发Solr大牛不太可能没有开发这个功能),功夫不负有心人,花了1天时间终于找到了完美的解决方法。
先说一下Solr的过滤机制是怎么实现的:
1、查询时将参数传入Request,经过处理,会调用SolrCore.execute()方法。
2、SolrCore.execute()会调用RequestHandlerBase.handleRequest()方法。
3、RequestHandlerBase.handleRequest()会调用Component,默认有6个Component。分别是QueryComponent、FacetComponent、MoreLikeThisComponent、HighlightComponent、StatsComponent、DebugComponent
4、fq是在QueryComponent解析的,
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
|
String[] fqs = req.getParams().getParams(CommonParams.FQ);
if
(fqs!=
null
&& fqs.length!=
0
) {
List<Query> filters = rb.getFilters();
if
(filters==
null
) {
filters =
new
ArrayList<Query>(fqs.length);
}
for
(String fq : fqs) {
if
(fq !=
null
&& fq.trim().length()!=
0
) {
QParser fqp = QParser.getParser(fq,
null
, req);
filters.add(fqp.getQuery());
}
}
// only set the filters if they are not empty otherwise
// fq=&someotherParam= will trigger all docs filter for every request
// if filter cache is disabled
if
(!filters.isEmpty()) {
rb.setFilters( filters );
}
}
|
实现对fq的解析,往里跟踪发现,是通过QueryParser.parse()将过滤条件转换成Query的。这个就是问题关键,既然filed:value形式的过滤条件可以实现,那filed:{value1 or value2}不就可以实现我想要的多值并集试的过滤吗?(QueryParser.parse()支持查询语法).
问题又来了,通过多次测试,发现value不能支持写查询表达式(纠结啊)。再一次耐心把代码看一下,突然发现一段奇怪的代码,QueryParsing.parseLocalParams()方法
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
|
int
off = start;
if
(!txt.startsWith(LOCALPARAM_START, off))
return
start;
StrParser p =
new
StrParser(txt, start, txt.length());
p.pos +=
2
;
// skip over "{!"
for
(; ; ) {
/*
if (p.pos>=txt.length()) {
throw new ParseException("Missing '}' parsing local params '" + txt + '"');
}
*/
char
ch = p.peek();
if
(ch == LOCALPARAM_END) {
return
p.pos +
1
;
}
String id = p.getId();
if
(id.length() ==
0
) {
throw
new
ParseException(
"Expected identifier '}' parsing local params '"
+ txt + '"');
}
String val =
null
;
ch = p.peek();
if
(ch !=
'='
) {
// single word... treat {!func} as type=func for easy lookup
val = id;
id = TYPE;
}
else
{
// saw equals, so read value
p.pos++;
ch = p.peek();
boolean
deref =
false
;
if
(ch ==
'$'
) {
p.pos++;
ch = p.peek();
deref =
true
;
// dereference whatever value is read by treating it as a variable name
}
if
(ch ==
'\"'
|| ch ==
'\''
) {
val = p.getQuotedString();
}
else
{
// read unquoted literal ended by whitespace or '}'
// there is no escaping.
int
valStart = p.pos;
for
(; ; ) {
if
(p.pos >= p.end) {
throw
new
ParseException(
"Missing end to unquoted value starting at "
+ valStart +
" str='"
+ txt +
"'"
);
}
char
c = p.val.charAt(p.pos);
if
(c == LOCALPARAM_END || Character.isWhitespace(c)) {
val = p.val.substring(valStart, p.pos);
break
;
}
p.pos++;
}
}
if
(deref) {
// dereference parameter
if
(params !=
null
) {
val = params.get(val);
}
}
}
if
(target !=
null
) target.put(id, val);
}
|
仔细看了一下,发现用于解析类似{!dismax v=$auth}这样的字符传(一直都不太想看这段代码,太长了),作用是将$auth替换为传入solr的参数,解析为{!dismax v=field4:2},后用dismax转换为Query,添加到过滤查询里。
最终解决方法是:索引
A{filter1:a,field2:a,field3:a,field4:1}
B{filter1:b,field2:b,field3:b,field4:2}
C{filter1:c,field2:c,field3:c,field4:3}
1
2
3
4
5
|
Map<String,Object> paramMap =
new
HashMap<String,Object>();
SearchParams params =
new
SearchParams(paramMap) ;
paramMap.put(
"auth"
,
"filed4:1 filed4:3"
);
//将过滤查询语句放入参数里
query.add(params);
query.addFilterQuery(
"{!dismax v=$auth}"
);
//通过dismax转换为query,加入查询语句,大功告成
|
结果:
A{filter1:a,field2:a,field3:a,field4:1}
C{filter1:c,field2:c,field3:c,field4:3}
附录
文中使用类的全路径
1
2
3
4
5
6
7
8
9
10
11
12
13
|
org.apache.solr.core.SolrCore
org.apache.solr.handler.StandardRequestHandler
org.apache.solr.handler.RequestHandlerBase.handleRequest
org.apache.solr.handler.StandardRequestHandler
org.apache.solr.handler.component.SearchHandler.handleRequestBody
org.apache.solr.handler.component.QueryComponent
org.apache.solr.handler.component.FacetComponent
org.apache.solr.handler.component.MoreLikeThisComponent
org.apache.solr.handler.component.HighlightComponent
org.apache.solr.handler.component.StatsComponent
org.apache.solr.handler.component.DebugComponent
org.apache.lucene.queryParser.QueryParser
org.apache.solr.search.QueryParsing
|