group by使用疑惑，有使用经验的分享一下

现遇到的问题：
不使用groupby 'tid'的结构集正常，我的目的是为了去重tid 使之唯一性
Array
(
    [error] =>
    [warning] =>
    [status] => 0
    [fields] => Array
        (
            [0] => subject
            [1] => content
        )

    [attrs] => Array
        (
            [tid] => 1
            [fid] => 1
            [authorid] => 1
            [postdate] => 2
        )

    [matches] => Array
        (
            [0] => Array
                (
                    [id] => 442194
                    [weight] => 2
                    [attrs] => Array
                        (
                            [tid] => 47220
                            [fid] => 64
                            [authorid] => 6557
                            [postdate] => 1282093487
                        )

                )

            [1] => Array
                (
                    [id] => 436380
                    [weight] => 2
                    [attrs] => Array
                        (
                            [tid] => 45968
                            [fid] => 24
                            [authorid] => 19694
                            [postdate] => 1281936645
                        )

                )
          [2] => Array
                (
                    [id] => 427360
                    [weight] => 2
                    [attrs] => Array
                        (
                            [tid] => 45968
                            [fid] => 24
                            [authorid] => 2400
                            [postdate] => 1281582950
                        )

                )
         ............


使用group by tid以后的结果集始终只有1条记录

Array
(
    [error] =>
    [warning] =>
    [status] => 0
    [fields] => Array
        (
            [0] => subject
            [1] => content
        )

    [attrs] => Array
        (
            [tid] => 1
            [fid] => 1
            [authorid] => 1
            [postdate] => 2
            [@groupby] => 1
            [@count] => 1
        )

    [matches] => Array
        (
            [0] => Array
                (
                    [id] => 442194
                    [weight] => 2
                    [attrs] => Array
                        (
                            [tid] => 47220
                            [fid] => 64
                            [authorid] => 6557
                            [postdate] => 1282093487
                            [@groupby] => 19700101
                            [@count] => 675
                        )

                )

        )
看了手册 sphinx的group by等同于mysql 的group by 为什么出来的结果确是这样？

请给出你的具体操作代码。

以下代码是逻辑处理
$sphinx = $this->_getSphinx();
list($host,$port) = $this->_getConfig();
$sphinx->SetServer ( $host, $port );
$sphinx->SetConnectTimeout ( 1 );
$sphinx->SetMatchMode ( $this->_getMode($method,$q) );
$digest && $sphinx->SetFilter ('digest',array(1,2));
$fid && $sphinx->SetFilter ('fid',$fid,$exclude);
$authorids && $sphinx->SetFilter ('authorid',$authorids);
if($sch_timemin && $sch_timemax){
      $sphinx->SetFilterRange('postdate',$sch_timemin,$sch_timemax);
}
$groupby && $sphinx->SetGroupBy ( $groupby, $this->getGroup(), "@group desc" );
$sortby && $sphinx->SetSortMode ( ($asc=='DESC' ? SPH_SORT_ATTR_DESC : SPH_SORT_ATTR_ASC), $sortby );
$page = isset($_GET['page']) ? $_GET['page'] : 1;
$page = max(1, intval($page));
$start_limit = intval(($page - 1) * $db_perpage);
$sphinx->SetLimits ( $start_limit, intval($db_perpage), ( $db_perpage>1000 ) ? $db_perpage : 1000 );
$sphinx->SetRankingMode ( $this->getRanking() );
$sphinx->SetArrayResult ( true );
$index = $this->getIndex($index);
$result = $sphinx->Query ( str_replace('|',' ',$q), $index );
跟踪到内核的Query函数 ->AddQuery 函数里的
的$req .= pack ( "NN", $this->_groupfunc, strlen($this->_groupby) ) . $this->_groupby;代码值都没什么问题
$this->_groupfunc = SPH_GROUPBY_ATTR
$this->_groupby = tid
以及$this->_groupsort = @group desc

在分析下去就碰内核了。sphinx group by的作用是否真的可以这样使用的？还是我会错了他的意？如果有使用过的人能否贴一下你的成功使用后代码，当然conf最好也贴一下。

1. 不使用group，返回结果有total_found属性吗？
2. 使用之后，还有吗？

total_found 使用后和使用前都有
但是使用后的值就变为1了
使用前应该是总记录数结果

我建议先直接写最简单的代码，把你的其他可能导致错误的逻辑都去掉。

去掉排序、翻页与限制，仅查询和分组。

this->sphinx->SetGroupBy( "class", SPH_GROUPBY_ATTR, "@group desc" );
$res = $this->sphinx->Query($this->query, 'all');
print_r($res);

我的就这么简单的查询，分组后出现的也是这个现象，请教下原因。手册我看了N++次了。。

同上描述：
total_found 使用后和使用前都有
但是使用SetGroupBy后的值就变为1了
使用前应该是总记录数结果

使用我们提供的下载演示程序，实际测试结果如下：
//$cl->SetGroupBy( "author_id", SPH_GROUPBY_ATTR, "@group desc" );
$res = $cl->Query ( '一个', "*" );
//print_r($cl);
print_r($res);

    [matches] => Array
        (
            [0] => Array
                (
                    [id] => 3
                    [weight] => 1319
                    [attrs] => Array
                        (
                            [published] => 1270094460
                            [author_id] => 2
                        )

                )

            [1] => Array
                (
                    [id] => 1
                    [weight] => 1252
                    [attrs] => Array
                        (
                            [published] => 1270131607
                            [author_id] => 1
                        )

                )

            [2] => Array
                (
                    [id] => 2
                    [weight] => 1252
                    [attrs] => Array
                        (
                            [published] => 1270135548
                            [author_id] => 1
                        )

                )

        )

    [total] => 3
    [total_found] => 3
    [time] => 0.001
    [words] => Array
        (
            [一个] => Array
                (
                    [docs] => 3
                    [hits] => 5
                )

        )

使用我们提供的下载演示程序，实际测试结果如下：
$cl->SetGroupBy( "author_id", SPH_GROUPBY_ATTR, "@group desc" );
$res = $cl->Query ( '一个', "*" );
//print_r($cl);
print_r($res);

    [matches] => Array
        (
            [0] => Array
                (
                    [id] => 3
                    [weight] => 1319
                    [attrs] => Array
                        (
                            [published] => 1270094460
                            [author_id] => 2
                            [@groupby] => 2
                            [@count] => 1
                        )

                )

            [1] => Array
                (
                    [id] => 1
                    [weight] => 1252
                    [attrs] => Array
                        (
                            [published] => 1270131607
                            [author_id] => 1
                            [@groupby] => 1
                            [@count] => 2
                        )

                )

        )

    [total] => 2
    [total_found] => 2
    [time] => 0.002
    [words] => Array
        (
            [一个] => Array
                (
                    [docs] => 3
                    [hits] => 5
                )

        )

$result = array();
                $result['words'] = $res['words'];
                if (is_array($res['matches']))
                {
                    //搜索结果的文档数目
                    $result['total'] = intval($res['total']);
                    $result['total_found'] = intval($res['total_found']);
                    //搜索用时
                    $result['time'] = $res['time'];
                    //分词情况和每个词的命中信息 文档数目
                    //存放搜索结果  每个元素为文档id 文档权重
                    $result['result'] = array();
                    //存放分组结果
                    $result['groupinfo'] = array();
                    foreach ( $res['matches'] as $docinfo )
                    {
                        //获取文档id和权重
                        $result['result'][$docinfo[id]] = $docinfo[weight];
                        //获取分组值和分组数目
                        $result['groupinfo'][$docinfo['attrs']['@groupby']] = $docinfo['attrs']['@count'];
                     }
                     return $result;
                }
                else
                {
                    //这个是没有找到结果的情况
                    $result['total'] = 0;
                    return $result;
                }