从Excel文件中读取所需类型的熊猫

Question

I am reading excel file using pandas containing 2 columns: df

EID      List of Tuples
 1       [('Physics', 90)]
 2       [('Physics', 80), ('Math', 70)]
 3       [('Physics', 60, ('Math', 25))]

when I check df['List of Tuples'].iat[0] it gives me u"[('Physics', 90)]" I am getting this as a unicode and not as a list of tuples. When I df['List of Tuples'].iat[0].decode('iso-8859-1').encode('utf-8') , I get string: "[('Physics', 90)]" I want to read/convert it as list of tuples [('Physics', 90)] instead of string or unicode.In short,I want to get rid of double quotes around each entry and read it as [('Physics', 90)] and [('Physics', 80), ('Math', 70)] and so on.

Answer 1

You might find it useful to parse these into python objects using ast which can convert string representations back into python objectd. Try something like the following (I can't replicate exactly because I don't have your data):

import ast
df['transformed_tuples'] = df['List of Tuples'].apply(ast.literal_eval)

To avoid this arising in the first place you might consider the file format you choose to read/write to, for example pickle will retain the original information (I'm assuming this has come from a pandas DataFrame that has been saved to excel).

You might also consider a tabular schema which doesn't have this irregular data type within it which would probably prove to be more stable and effective in the long run.