问题描述

我正在尝试使用代码将数据帧转换为系列，简化后如下所示:

Im attempting to convert a dataframe into a series using code which, simplified, looks like this:

dates = ['2016-1-{}'.format(i)for i in range(1,21)]
values = [i for i in range(20)]
data = {'Date': dates, 'Value': values}
df = pd.DataFrame(data)
df['Date'] = pd.to_datetime(df['Date'])
ts = pd.Series(df['Value'], index=df['Date'])
print(ts)

但是，打印输出如下所示:

However, print output looks like this:

Date
2016-01-01   NaN
2016-01-02   NaN
2016-01-03   NaN
2016-01-04   NaN
2016-01-05   NaN
2016-01-06   NaN
2016-01-07   NaN
2016-01-08   NaN
2016-01-09   NaN
2016-01-10   NaN
2016-01-11   NaN
2016-01-12   NaN
2016-01-13   NaN
2016-01-14   NaN
2016-01-15   NaN
2016-01-16   NaN
2016-01-17   NaN
2016-01-18   NaN
2016-01-19   NaN
2016-01-20   NaN
Name: Value, dtype: float64

NaN 是从哪里来的?DataFrame 对象上的视图是否不是 Series 类的有效输入?

Where does NaN come from? Is a view on a DataFrame object not a valid input for the Series class ?

我为 pd.Index 对象找到了 to_series 函数，DataFrames 有类似的东西吗?

I have found the to_series function for pd.Index objects, is there something similar for DataFrames ?

推荐答案

我觉得你可以使用 values，它将列 Value 转换为数组:

I think you can use values, it convert column Value to array:

ts = pd.Series(df['Value'].values, index=df['Date'])

import pandas as pd
import numpy as np
import io

dates = ['2016-1-{}'.format(i)for i in range(1,21)]
values = [i for i in range(20)]
data = {'Date': dates, 'Value': values}
df = pd.DataFrame(data)
df['Date'] = pd.to_datetime(df['Date'])
print df['Value'].values
[ 0  1  2  3  4  5  6  7  8  9 10 11 12 13 14 15 16 17 18 19]

ts = pd.Series(df['Value'].values, index=df['Date'])

print(ts)
Date
2016-01-01     0
2016-01-02     1
2016-01-03     2
2016-01-04     3
2016-01-05     4
2016-01-06     5
2016-01-07     6
2016-01-08     7
2016-01-09     8
2016-01-10     9
2016-01-11    10
2016-01-12    11
2016-01-13    12
2016-01-14    13
2016-01-15    14
2016-01-16    15
2016-01-17    16
2016-01-18    17
2016-01-19    18
2016-01-20    19
dtype: int64

或者你可以使用:

ts1 = pd.Series(data=values, index=pd.to_datetime(dates))
print(ts1)
2016-01-01     0
2016-01-02     1
2016-01-03     2
2016-01-04     3
2016-01-05     4
2016-01-06     5
2016-01-07     6
2016-01-08     7
2016-01-09     8
2016-01-10     9
2016-01-11    10
2016-01-12    11
2016-01-13    12
2016-01-14    13
2016-01-15    14
2016-01-16    15
2016-01-17    16
2016-01-18    17
2016-01-19    18
2016-01-20    19
dtype: int64

谢谢@ajcr 更好地解释为什么你得到 NaN:

Thank you @ajcr for better explanation why you get NaN:

当您将 Series 或 DataFrame 列提供给 pd.Series 时，它将使用 index 你指定.由于您的 DataFrame 列有一个整数 index(不是 date index)，因此您会得到很多缺失值.

When you give a Series or DataFrame column to pd.Series, it will reindex it using the index you specify. Since your DataFrame column has an integer index (not a date index) you get lots of missing values.