sphinx 同时使用多个索引进行检索探究

2014年2月15日 11:24:34

结论:

1.一次性使用多个索引进行查询的时候,返回的结果集中的fields字段没有什么清楚的意义(也没有找到文档对它的说明)

2.如果程序中一次搜索使用了多个索引,如果它们配置文件中过滤用的属性(aql_attr_uint,sql_field_string...)不全相同,那么最终返回的结果集中,只包含这几个索引中共有的属性

实验:

建立两个索引:goods_brand,  goods_cate, 分别是商品信息+品牌信息,商品信息+分类信息

 1  sql_query = select gid, gid as goodsid, siteid, catename from v_goods_info_cate

 2  sql_attr_uint = siteid

 3  sql_attr_uint = goodsid

 4  sql_field_string = catename

 5 

 6 #######################

 7 

 8  sql_query = select gid, gid as goodsid, siteid, brandname from v_goods_info_brand

 9  sql_attr_uint = siteid

10  sql_attr_uint = goodsid

11  sql_field_string = brandname

注:

1. brandname 是商品的品牌名字, catename是商品的分类名字

2. brandname, catename 在索引时,既作为全文索引,又作为属性值返回

3. siteid在两个索引中都有,brandname和catename只在各自的索引中存在

测试程序代码

1 $sphObj->AddQuery($keyword, 'goods_brand');

2 $sphObj->AddQuery($keyword, 'goods_cate');

3 $sphObj->AddQuery($keyword, 'goods_cate, goods_brand');

4 $sphObj->AddQuery($keyword, 'goods_brand,goods_cate');

5 

6 var_dump($rs[0]['fields'], $rs[0]['words'], $rs[0]['matches']);

注:

在程序中做控制:搜索"机"这个字,在goods_cate和goods_brand索引中各只有两条记录符合要求(一共有4条记录):

1.分别执行测试代码的第1行和第2行,并用第6行打印出结果:

 1 //goods_brand

 2 array (size=1)

 3   0 => string 'brandname' (length=9)

 4 

 5 array (size=1)

 6   '机' => 

 7     array (size=2)

 8       'docs' => string '10049' (length=5)

 9       'hits' => string '10049' (length=5)

10 

11 array (size=2)

12   0 => 

13     array (size=3)

14       'id' => string '157978' (length=6)

15       'weight' => string '1' (length=1)

16       'attrs' => 

17         array (size=3)

18           'goodsid' => string '157978' (length=6)

19           'siteid' => string '102' (length=3)

20           'brandname' => string '无锡一机' (length=12)

21   1 => 

22     array (size=3)

23       'id' => string '157980' (length=6)

24       'weight' => string '1' (length=1)

25       'attrs' => 

26         array (size=3)

27           'goodsid' => string '157980' (length=6)

28           'siteid' => string '102' (length=3)

29           'brandname' => string '无锡一机' (length=12)

30 

31 //goods_cate

32 array (size=1)

33   0 => string 'catename' (length=8)

34 

35 array (size=1)

36   '机' => 

37     array (size=2)

38       'docs' => string '43986' (length=5)

39       'hits' => string '43986' (length=5)

40 

41 array (size=2)

42   0 => 

43     array (size=3)

44       'id' => string '158010' (length=6)

45       'weight' => string '1' (length=1)

46       'attrs' => 

47         array (size=3)

48           'goodsid' => string '158010' (length=6)

49           'siteid' => string '102' (length=3)

50           'catename' => string '磨齿机' (length=9)

51   1 => 

52     array (size=3)

53       'id' => string '158014' (length=6)

54       'weight' => string '1' (length=1)

55       'attrs' => 

56         array (size=3)

57           'goodsid' => string '158014' (length=6)

58           'siteid' => string '102' (length=3)

59           'catename' => string '旋压机' (length=9)

View Code

注:

每个索引单独被使用时,各对应两条记录(一共有4条记录)

每条匹配记录中的'attrs'中有siteid+brandname,或者,siteid+catename

 

2.当用一次查询用多个索引时:分别执行第3行和第4行,并用第6行打印出结果:

 1 //goods_brand在前,goods_cate在后

 2 array (size=1)

 3   0 => string 'brandname' (length=9)

 4 

 5 array (size=1)

 6   '机' => 

 7     array (size=2)

 8       'docs' => string '54035' (length=5)

 9       'hits' => string '54035' (length=5)

10 

11 array (size=4)

12   0 => 

13     array (size=3)

14       'id' => string '157978' (length=6)

15       'weight' => string '1' (length=1)

16       'attrs' => 

17         array (size=2)

18           'goodsid' => string '157978' (length=6)

19           'siteid' => string '102' (length=3)

20   1 => 

21     array (size=3)

22       'id' => string '157980' (length=6)

23       'weight' => string '1' (length=1)

24       'attrs' => 

25         array (size=2)

26           'goodsid' => string '157980' (length=6)

27           'siteid' => string '102' (length=3)

28   2 => 

29     array (size=3)

30       'id' => string '158010' (length=6)

31       'weight' => string '1' (length=1)

32       'attrs' => 

33         array (size=2)

34           'goodsid' => string '158010' (length=6)

35           'siteid' => string '102' (length=3)

36   3 => 

37     array (size=3)

38       'id' => string '158014' (length=6)

39       'weight' => string '1' (length=1)

40       'attrs' => 

41         array (size=2)

42           'goodsid' => string '158014' (length=6)

43           'siteid' => string '102' (length=3)

44 

45 //goods_cate在前,goods_brand在后

46 array (size=1)

47   0 => string 'catename' (length=8)

48 

49 array (size=1)

50   '机' => 

51     array (size=2)

52       'docs' => string '54035' (length=5)

53       'hits' => string '54035' (length=5)

54 

55 array (size=4)

56   0 => 

57     array (size=3)

58       'id' => string '157978' (length=6)

59       'weight' => string '1' (length=1)

60       'attrs' => 

61         array (size=2)

62           'goodsid' => string '157978' (length=6)

63           'siteid' => string '102' (length=3)

64   1 => 

65     array (size=3)

66       'id' => string '157980' (length=6)

67       'weight' => string '1' (length=1)

68       'attrs' => 

69         array (size=2)

70           'goodsid' => string '157980' (length=6)

71           'siteid' => string '102' (length=3)

72   2 => 

73     array (size=3)

74       'id' => string '158010' (length=6)

75       'weight' => string '1' (length=1)

76       'attrs' => 

77         array (size=2)

78           'goodsid' => string '158010' (length=6)

79           'siteid' => string '102' (length=3)

80   3 => 

81     array (size=3)

82       'id' => string '158014' (length=6)

83       'weight' => string '1' (length=1)

84       'attrs' => 

85         array (size=2)

86           'goodsid' => string '158014' (length=6)

87           'siteid' => string '102' (length=3)

View Code

 注:

两个索引被同时使用,只有先后顺序不一样时,4条记录都得到了(这样的结果是对的)

但是第3行和第47行的代码键值对表明,返回的结果集中的fields值没有什么特别的含义(至少我不知到,难道只和排在前边的索引使用的全文索引字段同步?肯定有什么意义,只是我没有总结到吧)

另外,查看结果知道,每一条匹配记录的'attrs'数组中只有siteid键值对