Groupby作为具有MultiIndex

Question

我有一个类似的数据框

df = pd.DataFrame(
    pd.np.random.randint(1, 10, (20, 2)),
    index=["a", "b", "c", "d"] * 5,
    columns=["foo", "bar"],
)

看起来像这样

我正在尝试将其转换为以下多索引形式：4个轴标签中的每个标签都有两列，每个列均带有一个“ foo”和“ bar”子列：

我尝试摆弄df.groupby(df.index)，但后来我不知道如何将这些组变成列。

Answer 1

将GroupBy.cumcount用于DataFrame.set_index的索引中的if($order == '' ) { $order= id; } $homework6 = $mysqli->query("SELECT * FROM homework6 WHERE manufacturer_hebrew LIKE '%$x%' order by $order asc"); $rows = $homework6->fetch_all(MYSQLI_ASSOC); echo '<div >'; echo '<table border= "1" >'; echo '<th colspan="1">'. '<a href="?order=id">x</a>' .'</th>'; echo '<th colspan="1">'. '<a href="?order=manufacturer_hebrew">x</a>' .'</th>'; echo '<th colspan="1">'. '<a href="?order=manufacturer_english">x</a>' .'</th>'; echo '<th colspan="1">'. '<a href="?order=Models_number">x</a>' .'</th>'; echo '<th colspan="1">'. '<a href="?order=made_in">x</a>' .'</th>'; foreach ($rows as $paz) { if ($paz['Models_number'] < 2 ) { $color = "bg-danger"; } elseif ($paz['Models_number'] < 10){ $color = "bg-warning"; } elseif ($paz['Models_number'] < 15){ $color = "bg-success"; } else {$color = "bg-primary";} echo '<tr class='.$color.'>'; echo '<td >' .$paz['id']. '</td>'; echo '<td>' .$paz['manufacturer_hebrew']. '</td>'; echo '<td>' .$paz['manufacturer_english']. '</td>'; echo '<td>' .$paz['Models_number']. '</td>'; echo '<td >' .$paz['made_in']. '</td>'; echo '</tr>'; } echo '</table>'; echo '</div>'; ，然后通过DataFrame.unstack进行整形，并使用DataFrame.swaplevel来更改DataFrame.sort_index的列中MultiIndex的顺序Boost Yap：

Multiindex