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
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
|
<test>
<name>idf=plain vs distributed</name>
<config>
indexer
{
mem_limit = 16M
}
searchd
{
<searchd_settings/>
workers = threads
}
source src_base
{
type = mysql
<sql_settings/>
}
source src_1 : src_base
{
sql_query = SELECT * FROM test_table WHERE gid=1
sql_attr_uint = gid
}
source src_2 : src_1
{
sql_query = SELECT * FROM test_table WHERE gid=2
sql_attr_uint = gid
}
index i1
{
source = src_1
path = <data_path/>/i1
docinfo = extern
}
index i2
{
source = src_2
path = <data_path/>/i2
docinfo = extern
}
index dd
{
type = distributed
agent = <my_address/>:i1
agent = <my_address/>:i2
}
source src_kw : src_1
{
sql_query = SELECT 1 as idd, 11 as gid, 'test starts here' as text UNION SELECT 2 as idd, 11 as gid, 'the text from hox' as text UNION SELECT 3 as idd, 11 as gid, 'from moon to fox' as text UNION SELECT 4 as idd, 11 as gid, 'fox from the box' as text UNION SELECT 5 as idd, 11 as gid, 'home no here' as text
sql_attr_uint = gid
}
index kw
{
source = src_kw
path = <data_path/>/kw
docinfo = extern
min_prefix_len = 1
dict = keywords
}
source src_l1 : src_base
{
sql_query = SELECT * FROM test_table WHERE gid=3
sql_attr_uint = gid
}
source src_l2 : src_1
{
sql_query = SELECT * FROM test_table WHERE gid=4
sql_attr_uint = gid
}
index l1
{
source = src_l1
path = <data_path/>/l1
docinfo = extern
}
index l2
{
source = src_l2
path = <data_path/>/l2
docinfo = extern
}
index dl
{
type = distributed
agent = <my_address/>:l1,l2
}
</config>
<db_create>
CREATE TABLE test_table
(
id INT NOT NULL,
gid INT NOT NULL,
text VARCHAR(255) NOT NULL DEFAULT ''
);
</db_create>
<db_drop>
DROP TABLE IF EXISTS test_table;
</db_drop>
<db_insert>
INSERT INTO test_table VALUES
( 1, 1, 'da one' ),
( 2, 1, 'da two' ),
( 3, 1, 'da three' ),
( 11, 2, 'da blow' ),
( 12, 2, 'da pills' ),
( 13, 2, 'da yak' ),
( 14, 2, 'da herb' ),
( 15, 2, 'da blow' ),
( 100, 3, 'da blow' ),
( 101, 3, 'da win' ),
( 102, 3, 'da yak' ),
( 103, 3, 'da herb' ),
( 104, 3, 'da blow' ),
( 200, 4, 'da white' ),
( 201, 4, 'da win' ),
( 202, 4, 'da win' ),
( 203, 4, 'da win' ),
( 204, 4, 'da blow' )
</db_insert>
<sphqueries>
<sphinxql>select *, weight() from dd where match('da')</sphinxql>
<sphinxql>select *, weight() from dd where match('da') option idf=plain</sphinxql>
<!-- regression idf options vs distributed index -->
<sphinxql>select *, packedfactors() as f, weight() from i1,i2 where match('da blow') option idf=plain,idf=tfidf_unnormalized, ranker=expr('sum(lcs*user_weight)*1000+bm25')</sphinxql>
<sphinxql>select *, packedfactors() as f, weight() from dd where match('da blow') option idf=plain,idf=tfidf_unnormalized, ranker=expr('sum(lcs*user_weight)*1000+bm25')</sphinxql>
<sphinxql>select *, packedfactors() as f, weight() from i1,i2 where match('da blow') option idf=plain,idf=tfidf_normalized, ranker=expr('sum(lcs*user_weight)*1000+bm25')</sphinxql>
<sphinxql>select *, packedfactors() as f, weight() from dd where match('da blow') option idf=plain,idf=tfidf_normalized, ranker=expr('sum(lcs*user_weight)*1000+bm25')</sphinxql>
<!-- regression expression ranker crash on expanded terms -->
<sphinxql>select *, weight() as w, packedfactors() as p from kw where match('t* here') option ranker=expr('sum(lcs)*1000+bm25')</sphinxql>
<sphinxql>select *, weight() as w, packedfactors() as p from kw where match('t* starts h*') option ranker=expr('sum(lcs)*1000+bm25')</sphinxql>
<sphinxql>select *, weight() as w, packedfactors() as p from kw where match('t* from h*') option ranker=expr('sum(lcs)*1000+bm25')</sphinxql>
<sphinxql>select *, weight() as w, packedfactors() as p from kw where match('t* ( starts | from ) h*') option ranker=expr('sum(lcs)*1000+bm25')</sphinxql>
<!-- regression local idf -->
<sphinxql>select *, weight() from l1,l2 where match('da win') option local_df=1</sphinxql>
<sphinxql>select *, weight() from l1,l2 where match('da win')</sphinxql>
<sphinxql>select *, weight() from dl where match('da win') option local_df=1</sphinxql>
<sphinxql>select *, weight() from dl where match('da win')</sphinxql>
<sphinxql>select *, weight() from l1,l2 where match('blow') option local_df=1</sphinxql>
<sphinxql>select *, weight() from l1,l2 where match('blow')</sphinxql>
<sphinxql>select *, weight() from dl where match('blow') option local_df=1</sphinxql>
<sphinxql>select *, weight() from dl where match('blow')</sphinxql>
<sphinxql>select id from kw order by weight() asc</sphinxql>
</sphqueries>
</test>
|
