问题
I'm trying to load a CSV file (this file) to create a multiindex e multilevel dataframe. It has 5(five) indexes and 3(three) levels in columns.
How I can do? Here is the code:
df = pd.read_csv('./teste.csv'
,index_col=[0,1,2,3,4]
,header=[0,1,2,3]
,skipinitialspace=True
,tupleize_cols=True)
df.columns = pd.MultiIndex.from_tuples(df.columns)
Expected output:
variables u \
level 1
days 1 2
times 00h 06h 12h 18h 00h
wsid lat lon start prcp_24
329 -43.969397 -19.883945 2007-03-18 10:00:00 72.0 0 0 0 0 0
2007-03-20 10:00:00 104.4 0 0 0 0 0
2007-10-18 23:00:00 92.8 0 0 0 0 0
2007-12-21 00:00:00 60.4 0 0 0 0 0
2008-01-19 18:00:00 53.0 0 0 0 0 0
2008-04-05 01:00:00 80.8 0 0 0 0 0
2008-10-31 17:00:00 101.8 0 0 0 0 0
2008-11-01 04:00:00 82.0 0 0 0 0 0
2008-12-29 00:00:00 57.8 0 0 0 0 0
2009-03-28 10:00:00 72.4 0 0 0 0 0
2009-10-07 02:00:00 57.8 0 0 0 0 0
2009-10-08 00:00:00 83.8 0 0 0 0 0
2009-11-28 16:00:00 84.4 0 0 0 0 0
2009-12-18 04:00:00 51.8 0 0 0 0 0
2009-12-28 00:00:00 96.4 0 0 0 0 0
2010-01-06 05:00:00 74.2 0 0 0 0 0
2011-12-18 00:00:00 113.6 0 0 0 0 0
2011-12-19 00:00:00 90.6 0 0 0 0 0
2012-11-15 07:00:00 85.8 0 0 0 0 0
2013-10-17 00:00:00 52.4 0 0 0 0 0
2014-04-01 22:00:00 72.0 0 0 0 0 0
2014-10-20 06:00:00 56.6 0 0 0 0 0
2014-12-13 09:00:00 104.4 0 0 0 0 0
2015-02-09 00:00:00 62.0 0 0 0 0 0
2015-02-16 19:00:00 56.8 0 0 0 0 0
2015-05-06 17:00:00 50.8 0 0 0 0 0
2016-02-26 00:00:00 52.2 0 0 0 0 0
I need handling error "TypeError: Expected tuple, got str":
TypeError: Expected tuple, got str
回答1:
You are getting an error because some of your columns are not tuples, they are strings from index 2368 to 2959 in df.columns.
Indices where the columns are strings:
df.columns[2368:2959]
Index(['('z', '1', '1', '00h').1', '('z', '1', '1', '06h').1',
'('z', '1', '1', '12h').1', '('z', '1', '1', '18h').1',
'('z', '1', '2', '00h').1', '('z', '1', '2', '06h').1',
'('z', '1', '2', '12h').1', '('z', '1', '2', '18h').1',
'('z', '1', '3', '00h').1', '('z', '1', '3', '06h').1',
...
'('z', '1000', '2', '06h').1', '('z', '1000', '2', '12h').1',
'('z', '1000', '2', '18h').1', '('z', '1000', '3', '00h').1',
'('z', '1000', '3', '06h').1', '('z', '1000', '3', '12h').1',
'('z', '1000', '3', '18h').1', '('z', '1000', '4', '00h').1',
'('z', '1000', '4', '06h').1', '('z', '1000', '4', '12h').1'],
dtype='object', length=591)
Since you want multi-index column dataframe using the tuples, so we are cleaning these strings first by taking the substring which is necessary using re.findall with regex pattern = '(\(.*?\)).' then passing this value through ast.literal_eval for converting string to tuple automatically. Finally, using the pd.MultiIndex.from_tuples as:
df = pd.read_csv('teste.csv',index_col=[0,1,2,3,4],header=[0,1,2,3],parse_dates=True)
import re
import ast
column_list = []
for column in df.columns:
if isinstance(column,str):
column_list.append(ast.literal_eval(re.findall('(\(.*?\)).',column)[0]))
else:
column_list.append(column)
df.columns = pd.MultiIndex.from_tuples(column_list,
names=('variables', 'level','days','times'))
print(df.iloc[:,:6].head())
variables u
level 1
days 1 2
times 00h 06h 12h 18h 00h 06h
wsid lat lon start prcp_24
329 -43.969397 -19.883945 2007-03-18 10:00:00 72.0 0 0 0 0 0 0
2007-03-20 10:00:00 104.4 0 0 0 0 0 0
2007-10-18 23:00:00 92.8 0 0 0 0 0 0
2007-12-21 00:00:00 60.4 0 0 0 0 0 0
2008-01-19 18:00:00 53.0 0 0 0 0 0 0
来源：https://stackoverflow.com/questions/53022580/handling-error-typeerror-expected-tuple-got-str-loading-a-csv-to-pandas-mult