Skip to content

Commit

Permalink
add files
Browse files Browse the repository at this point in the history
  • Loading branch information
继盛 committed Feb 1, 2016
1 parent d8a1547 commit 58cd662
Show file tree
Hide file tree
Showing 90 changed files with 24,913 additions and 0 deletions.
Binary file added docs/source/_static/output_10_1.png
View file
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Binary file added docs/source/_static/output_29_1.png
View file
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Binary file added docs/source/_static/output_31_1.png
View file
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Binary file added docs/source/_static/output_7_1.png
View file
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Binary file added docs/source/_static/output_8_1.png
View file
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
Binary file added docs/source/_static/output_9_1.png
View file
Sorry, something went wrong. Reload?
Sorry, we cannot display this file.
Sorry, this file is invalid so it cannot be displayed.
427 changes: 427 additions & 0 deletions docs/source/cl-zh.rst
View file

Large diffs are not rendered by default.

348 changes: 348 additions & 0 deletions docs/source/df-agg-zh.rst
View file
Original file line number Diff line number Diff line change
@@ -0,0 +1,348 @@
.. _dfagg:

.. code:: python
from odps.df import DataFrame
.. code:: python
iris = DataFrame(o.get_table('pyodps_iris'))
聚合操作
========

首先，我们可以使用\ ``describe``\ 函数，来查看DataFrame里数字列的数量、最大值、最小值、平均值以及标准差是多少。

.. code:: python
print(iris.describe())
.. code:: python
sepallength_count sepallength_min sepallength_max sepallength_mean \
0 150 4.3 7.9 5.843333
sepallength_std sepalwidth_count sepalwidth_min sepalwidth_max \
0 0.825301 150 2 4.4
sepalwidth_mean sepalwidth_std petallength_count petallength_min \
0 3.054 0.432147 150 1
petallength_max petallength_mean petallength_std petalwidth_count \
0 6.9 3.758667 1.758529 150
petalwidth_min petalwidth_max petalwidth_mean petalwidth_std
0 0.1 2.5 1.198667 0.760613
我们可以使用单列来执行聚合操作：

.. code:: python
iris.sepallength.max()
.. code:: python
7.9
支持的聚合操作包括：

.. raw:: html

<div style='padding-bottom: 30px'>
<table border="1" class="dataframe">
<tr>
<th>聚合操作</th>
<th>说明</th>
</tr>
<tr>
<td>count（或size）</td>
<td>数量</td>
</tr>
<tr>
<td>min</td>
<td>最小值</td>
</tr>
<tr>
<td>max</td>
<td>最大值</td>
</tr>
<tr>
<td>sum</td>
<td>求和</td>
</tr>
<tr>
<td>mean</td>
<td>均值</td>
</tr>
<tr>
<td>median</td>
<td>中位数</td>
</tr>
<tr>
<td>var</td>
<td>方差</td>
</tr>
<tr>
<td>std</td>
<td>标准差</td>
</tr>
</table>
</div>

分组聚合
========

DataFrame
API提供了groupby来执行分组操作，分组后的一个主要操作就是通过调用agg或者aggregate方法，来执行聚合操作。

.. code:: python
iris.groupby('name').agg(iris.sepallength.max(), smin=iris.sepallength.min())
.. raw:: html

<div style='padding-bottom: 30px'>
<table border="1" class="dataframe">
<thead>
<tr style="text-align: right;">
<th></th>
<th>name</th>
<th>sepallength_max</th>
<th>smin</th>
</tr>
</thead>
<tbody>
<tr>
<th>0</th>
<td>Iris-setosa</td>
<td>5.8</td>
<td>4.3</td>
</tr>
<tr>
<th>1</th>
<td>Iris-versicolor</td>
<td>7.0</td>
<td>4.9</td>
</tr>
<tr>
<th>2</th>
<td>Iris-virginica</td>
<td>7.9</td>
<td>4.9</td>
</tr>
</tbody>
</table>
</div>



最终的结果列中会包含分组的列，以及聚合的列。

DataFrame
API提供了一个\ ``value_counts``\ 操作，能返回按某列分组后，每个组的个数从大到小排列的操作。

我们使用groupby表达式可以写成：

.. code:: python
iris.groupby('name').agg(count=iris.name.count()).sort('count', ascending=False).head(5)
.. raw:: html

<div style='padding-bottom: 30px'>
<table border="1" class="dataframe">
<thead>
<tr style="text-align: right;">
<th></th>
<th>name</th>
<th>count</th>
</tr>
</thead>
<tbody>
<tr>
<th>0</th>
<td>Iris-virginica</td>
<td>50</td>
</tr>
<tr>
<th>1</th>
<td>Iris-versicolor</td>
<td>50</td>
</tr>
<tr>
<th>2</th>
<td>Iris-setosa</td>
<td>50</td>
</tr>
</tbody>
</table>
</div>



使用value\_counts就很简单了：

.. code:: python
iris['name'].value_counts().head(5)
.. raw:: html

<div style='padding-bottom: 30px'>
<table border="1" class="dataframe">
<thead>
<tr style="text-align: right;">
<th></th>
<th>name</th>
<th>count</th>
</tr>
</thead>
<tbody>
<tr>
<th>0</th>
<td>Iris-virginica</td>
<td>50</td>
</tr>
<tr>
<th>1</th>
<td>Iris-versicolor</td>
<td>50</td>
</tr>
<tr>
<th>2</th>
<td>Iris-setosa</td>
<td>50</td>
</tr>
</tbody>
</table>
</div>



对于聚合后的单列操作，我们也可以直接取出列名。但此时只能使用聚合函数。

.. code:: python
iris.groupby('name').petallength.sum()
.. raw:: html

<div style='padding-bottom: 30px'>
<table border="1" class="dataframe">
<thead>
<tr style="text-align: right;">
<th></th>
<th>petallength_sum</th>
</tr>
</thead>
<tbody>
<tr>
<th>0</th>
<td>73.2</td>
</tr>
<tr>
<th>1</th>
<td>213.0</td>
</tr>
<tr>
<th>2</th>
<td>277.6</td>
</tr>
</tbody>
</table>
</div>



.. code:: python
iris.groupby('name').agg(iris.petallength.notnull().sum())
.. raw:: html

<div style='padding-bottom: 30px'>
<table border="1" class="dataframe">
<thead>
<tr style="text-align: right;">
<th></th>
<th>name</th>
<th>petallength_sum</th>
</tr>
</thead>
<tbody>
<tr>
<th>0</th>
<td>Iris-setosa</td>
<td>50</td>
</tr>
<tr>
<th>1</th>
<td>Iris-versicolor</td>
<td>50</td>
</tr>
<tr>
<th>2</th>
<td>Iris-virginica</td>
<td>50</td>
</tr>
</tbody>
</table>
</div>



分组时也支持对常量进行分组，但是需要使用Scalar初始化。

.. code:: python
from odps.df import Scalar
.. code:: python
iris.groupby(Scalar(1)).petallength.sum()
.. raw:: html

<div style='padding-bottom: 30px'>
<table border="1" class="dataframe">
<thead>
<tr style="text-align: right;">
<th></th>
<th>petallength_sum</th>
</tr>
</thead>
<tbody>
<tr>
<th>0</th>
<td>563.8</td>
</tr>
</tbody>
</table>
</div>


0 comments on commit 58cd662

Please sign in to comment.