python - 更改 pandas 0.13.0 "print dataframe"以像早期版本一样打印数据框

Question

在 pandas 的新版本 0.13.0 中，数据帧 df 打印在一长串数字中，使用

df

或

print df

而不是像以前那样的概述，现在只能使用

df.info()

是否可以更改默认的“df”或“print df”命令以显示:

In [12]: df.info()
<class 'pandas.core.frame.DataFrame'>
DatetimeIndex: 4319 entries, 2010-02-18 00:00:00 to 2010-03-13 23:15:00
Data columns (total 2 columns):
QInt    4319  non-null values
QHea    4319  non-null values
dtypes: float32(2)

再次代替:

In [11]: df
Out[11]:
                                  QInt         QHea
2010-02-18 00:00:00         169.666672     0.000000
2010-02-18 00:15:00         152.000000    -0.000000
2010-02-18 00:15:00         152.000000    -0.000000
2010-02-18 00:30:00         155.000000    -0.000000
2010-02-18 00:30:04         155.063950    -0.000000
2010-02-18 00:30:04         155.063950 -1136.823364
2010-02-18 00:45:00         169.666672  4587.430176
2010-02-18 01:00:00         137.333328  4532.890137
2010-02-18 01:00:00         137.333328  4532.890137
2010-02-18 01:15:00         177.000000  4464.479980
2010-02-18 01:15:00         177.000000  4464.479980
2010-02-18 01:30:00         169.666672  4391.839844
2010-02-18 01:30:00         169.666672  4391.839844
2010-02-18 01:45:00         155.000000  4313.049805
2010-02-18 01:45:00         155.000000  4313.049805
2010-02-18 02:00:00         144.666672  4230.100098
2010-02-18 02:15:00         162.333328  4144.819824
2010-02-18 02:15:00         162.333328  4144.819824
2010-02-18 02:30:00         177.000000  4059.689941
2010-02-18 02:45:00         144.666672  3987.149902
2010-02-18 02:45:00         144.666672  3987.149902
2010-02-18 03:00:00         155.000000  3924.629883
2010-02-18 03:00:00         155.000000  3924.629883
2010-02-18 03:15:00         162.333328  3865.129883
2010-02-18 03:15:00         162.333328  3865.129883
2010-02-18 03:30:00         162.333328  3811.050049
2010-02-18 03:30:00         162.333328  3811.050049
2010-02-18 03:45:00         152.000000  3765.590088
2010-02-18 03:45:00         152.000000  3765.590088
2010-02-18 04:00:00         162.333328  3735.080078
2010-02-18 04:15:00         162.333328  3703.169922
2010-02-18 04:15:00         162.333328  3703.169922
2010-02-18 04:30:00         144.666672  3673.139893
2010-02-18 04:45:00         169.666672  3647.100098
2010-02-18 04:45:00         169.666672  3647.100098
2010-02-18 05:00:00         162.333328  3622.129883
2010-02-18 05:15:00         155.000000  3594.159912
2010-02-18 05:15:00         155.000000  3594.159912
2010-02-18 05:30:00         159.333328  3569.699951
2010-02-18 05:30:00         159.333328  3569.699951
2010-02-18 05:45:00         147.666672  3551.179932
2010-02-18 05:45:00         147.666672  3551.179932
2010-02-18 06:00:00         177.000000  3531.669922
2010-02-18 06:00:00         177.000000  3531.669922
2010-02-18 06:15:00         159.333328  3514.679932
2010-02-18 06:15:00         159.333328  3514.679932
2010-02-18 06:30:00         155.000000  3499.669922
2010-02-18 06:30:00         155.000000  3499.669922
2010-02-18 06:45:00         155.000000  3485.320068
2010-02-18 06:45:00         155.000000  3485.320068
2010-02-18 06:59:54.750000  162.291245    19.999992
2010-02-18 06:59:54.750000  162.291245     0.000000
2010-02-18 07:00:00         162.333328     0.000000
2010-02-18 07:00:00         162.333328     0.000000
2010-02-18 07:15:00         166.666672     0.000000
2010-02-18 07:15:00         166.666672     0.000000
2010-02-18 07:30:00         155.000000     0.000000
2010-02-18 07:30:00         155.000000     0.000000
2010-02-18 07:45:00         155.000000     0.000000
2010-02-18 07:45:00         155.000000     0.000000
                                   ...          ...

[4319 rows x 2 columns]

Answer 1

设置

pd.options.display.large_repr = 'info'

自 v.0.13 起的默认值为“截断”。

In [93]: df = pd.DataFrame(np.arange(4319*2).reshape(4319,2))

In [94]: pd.options.display.large_repr = 'info'

In [95]: df
Out[95]: 
<class 'pandas.core.frame.DataFrame'>
Int64Index: 4319 entries, 0 to 4318
Data columns (total 2 columns):
0    4319 non-null int32
1    4319 non-null int32
dtypes: int32(2)

---

我通过在输出中搜索字符串 'info()' 找到了这个:

In [65]: pd.set_option?

---

要使其成为交互式 session 的默认行为:

如果你还没有设置它，定义环境变量PYTHONSTARTUP类似于 /home/user/bin/startup.py

然后编辑/创建 /home/user/bin/startup.py 以包含类似的内容

import pandas as pd
pd.options.display.large_repr = 'info'

现在，无论何时启动交互式 Python session ，startup.py 文件都会被执行，您将可以通过 pd 变量访问 pandas，并且 large_repr 默认为 'info'。