Question

这是一张包含product_id及其功能

的表格

| product_id | feature_title | feature_value |
+------------+---------------+---------------+
|    6312    |   warranty    |       1       |
|    6312    |     color     |      red      |
|    2083    |   warranty    |       1       |
|    2083    |     color     |     blue      |
|    8686    |   warranty    |       0       |
|    8686    |     color     |      red      |

如何仅选择具有这些细节的记录：

warranty = 1
color = red

Answer 1

我认为你需要这样的东西：

select *
from (
    select product_id
        ,max(case when feature_title = 'warranty' then feature_value end) warranty
        ,max(case when feature_title = 'color' then feature_value end) color
    from yourTable
    group by product_id           -- grouping to one row per id
   ) pivot
where
    warranty = '1' and
    color = 'red';

My SQL Fiddle Demo
SQL Server Fiddle Demo

当我想要比较具有相同id的多个记录时，我更喜欢将这些记录分组到一行，例如转动表格。使用max(case)转换数据的方法如上所示在内部选择中。<登记/> 例如case when feature_title = 'warranty' then feature_value end feature_value feature_title为'warranty'时feature_title的结果为null，max为feature_value feature_value id= mapall id=mapall2空值和var =markers and var = markers2将为<script> <?php echo "var markers=$markers;\n"; ?> function initMap() { var latlng = new google.maps.LatLng(-34.031342, 18.577419); // default location var myOptions = { zoom: 16, center: latlng, mapTypeId: google.maps.MapTypeId.SATELLITE, mapTypeControl: true }; var map = new google.maps.Map(document.getElementById('mapall'),myOptions); var infowindow = new google.maps.InfoWindow(), marker, lat, lng; for( i = 0; i < markers.length; i++ ) { lat = (markers[i].GPS1); lng = (markers[i].GPS2); name = (markers[i].client_address); marker = new google.maps.Marker({ position: new google.maps.LatLng(lat,lng), name:name, map: map }); google.maps.event.addListener( marker, 'click', function(e){ infowindow.setContent( this.name ); infowindow.open( map, this ); }.bind( marker ) ); } } </script>。 -HTH

Answer 2

我会使用聚合来做到这一点：

select product_id
from t
group by product_id
having sum(case when feature_title = 'warranty' and feature_value = '1' then 1 else 0 end) > 0 and
       sum(case when feature_title = 'color' and feature_value = 'red' then 1 else 0 end) > 0;

首先，并非所有数据库都支持pivot。其次，我不认为子查询会给查询增加太多。

注意：也许更简单的版本是：

select product_id
from t
where (feature_title = 'warranty' and feature_value = '1') or
      (feature_title = 'color' and feature_value = 'red')
group by product_id
having count(distinct feature_title) = 2;

SQL在同一列中选择具有不同值的记录

2 个答案: