Python 如何将 Pandas 数据框的数据类型更改为具有定义格式的字符串？
声明：本页面是StackOverFlow热门问题的中英对照翻译，遵循CC BY-SA 4.0协议，如果您需要使用它，必须同样遵循CC BY-SA许可，注明原文地址和作者信息，同时你必须将它归于原作者（不是我）：StackOverFlow
原文地址: http://stackoverflow.com/questions/22276503/
Warning: these are provided under cc-by-sa 4.0 license. You are free to use/share it, But you must attribute it to the original authors (not me):
StackOverFlow
How do I change data-type of pandas data frame to string with a defined format?
提问by user1718097
I'm starting to tear my hair out with this - so I hope someone can help. I have a pandas DataFrame that was created from an Excel spreadsheet using openpyxl. The resulting DataFrame looks like:
我开始用这个撕掉我的头发 - 所以我希望有人能帮忙。我有一个使用 openpyxl 从 Excel 电子表格创建的 Pandas DataFrame。生成的 DataFrame 如下所示：
print image_name_data
id image_name
0 1001 1001_mar2014_report
1 1002 1002_mar2014_report
2 1003 1003_mar2014_report
[3 rows x 2 columns]
…with the following datatypes:
...具有以下数据类型：
print image_name_data.dtypes
id float64
image_name object
dtype: object
The issue is that the numbers in the id column are, in fact, identification numbers and I need to treat them as strings. I've tried converting the id column to strings using:
问题是 id 列中的数字实际上是标识号，我需要将它们视为字符串。我尝试使用以下方法将 id 列转换为字符串：
image_name_data['id'] = image_name_data['id'].astype('str')
This seems a bit ugly but it does produce a variable of type 'object' rather than 'float64':
这看起来有点难看，但它确实产生了一个类型为“object”而不是“float64”的变量：
print image_name_data.dyptes
id object
image_name object
dtype: object
However, the strings that are created have a decimal point, as shown:
但是，创建的字符串有一个小数点，如下所示：
print image_name_data
id image_name
0 1001.0 1001_mar2014_report
1 1002.0 1002_mar2014_report
2 1003.0 1003_mar2014_report
[3 rows x 2 columns]
How can I convert a float64 column in a pandas DataFrame to a string with a given format (in this case, for example, '%10.0f')?
如何将 Pandas DataFrame 中的 float64 列转换为具有给定格式的字符串（在本例中，例如，'%10.0f'）？
采纳答案by exp1orer
I'm unable to reproduce your problem but have you tried converting it to an integer first?
我无法重现您的问题，但您是否尝试先将其转换为整数？
image_name_data['id'] = image_name_data['id'].astype(int).astype('str')
Then, regarding your more general question you could use map(as in this answer). In your case:
然后，关于您可以使用的更一般的问题map（如本答案所示）。在你的情况下：
image_name_data['id'] = image_name_data['id'].map('{:.0f}'.format)
回答by exp1orer
I'm putting this in a new answer because no linebreaks / codeblocks in comments. I assume you want those nans to turn into a blank string? I couldn't find a nice way to do this, only do the ugly method:
我将此放在一个新答案中，因为注释中没有换行符/代码块。我假设您希望这些 nans 变成空白字符串？我找不到一个很好的方法来做到这一点，只能做丑陋的方法：
s = pd.Series([1001.,1002.,None])
a = s.loc[s.isnull()].fillna('')
b = s.loc[s.notnull()].astype(int).astype(str)
result = pd.concat([a,b])
回答by smishra
If you could reload this, you might be able to use dtypes argument.
如果您可以重新加载它，您也许可以使用 dtypes 参数。
pd.read_csv(..., dtype={'COL_NAME':'str'})
