在 Pandas 中动态命名 DataFrame
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/32577939/
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
Dynamically naming DataFrames in Pandas
提问by Harrison Wallace
Imagine I have 2 data frames as such:
想象一下，我有 2 个数据框：
foo = pd.DataFrame({'a': [1,2,3], 'b': [4,5,6]})
bar = pd.DataFrame({'c':[7,8,9], 'd':[10,11,12]})
I want to subset each of these data frames and put them in a new data frame with a dynamic name. When I look up anything on dynamic naming in python, they say, don't do it, use a dictionary. I can't quite figure out how to make it work though. Essentially I want the following:
我想对这些数据帧中的每一个进行子集化，并将它们放入一个具有动态名称的新数据帧中。当我在 python 中查找有关动态命名的任何内容时，他们说，不要这样做，使用字典。我不太清楚如何使它工作。基本上我想要以下内容：
foo_first = foo[0:1]
bar_first = bar[0:1]
But I want to be able to do it looping through a list. I would imagine it something like this, if I'm trying to do it with a dictionary:
但我希望能够循环遍历列表。如果我想用字典来做，我会想象它是这样的：
dfs_list = [foo, bar]
dfs_dict = {}
for x in dfs_list:
dfs_dict[x+'_first']=foo[0:1]
Which does not work.
这不起作用。
You might be wondering what I'm actually trying to do, as my example is so arbitrary and pointless. In my real world example, I have several data frames indexed by date. I want to create new dataframes based on the names of these old dataframes for the current year and month. So if foo and bar were giant datasets with date indexes, I want to automate:
你可能想知道我到底在做什么，因为我的例子太武断和毫无意义。在我的真实示例中，我有几个按日期索引的数据框。我想根据当前年份和月份的这些旧数据框的名称创建新数据框。因此，如果 foo 和 bar 是带有日期索引的巨型数据集，我想自动化：
foo_year = foo['2015-01-01':'2015-12-31']
bar_year = bar['2015-01-01':'2015-12-31']
foo_month = foo['2015-08-01':'2015-08-31']
bar_month = foo['2015-08-01':'2015-08-31']
Any help?
有什么帮助吗？
回答by Blckknght
I can't think of any reason you couldn't use a dictionary of DataFrames. This will let you avoid needing to treat the variable names as data:
我想不出任何原因您不能使用DataFrames字典。这将使您无需将变量名称视为数据：
whole_dataframes = {"foo": foo, "bar": bar}
first_dataframes = {name: value[:1] for name, value in whole_dataframes.items()}
I'm using the fooand barvariables you've described to initialize the first dict, but you can probably skip that step and just create the values directly in the dict:
我正在使用您描述的foo和bar变量来初始化第一个 dict，但您可以跳过该步骤，直接在 dict 中创建值：
whole_dataframes = {}
whole_dataframes["foo"] = pd.DataFrame({'a': [1,2,3], 'b': [4,5,6]})
whole_dataframes["bar"] = pd.DataFrame({'c':[7,8,9], 'd':[10,11,12]})
