does anybody know how to remove the header of the index colum when applying pandas.read_csv and to_csv?
Example:
import datetime
import pandas as pd
from io import BytesIO
todays_date = datetime.datetime.now().date()
earlierDates = 2
index = pd.date_range(todays_date-datetime.timedelta(earlierDates), \
periods=earlierDates, freq='D')
columns = ['A', 'B']
df1 = pd.DataFrame(index=index, columns=columns)
df1.loc[index[0]] = [2, 'test1a test1b']
df1.loc[index[1]] = [1, 'test2a test2b']
df1
Out[85]:
A B
2018-08-09 2 test1a test1b
2018-08-10 1 test2a test2b
df2 = pd.read_csv(BytesIO(df1.to_csv()))
df2
Out[83]:
Unnamed: 0 A B
0 2018-08-09 2 test1a test1b
1 2018-08-10 1 test2a test2b
How can one remove the header "Unnamed: 0" in the above output?
Doing the following removes the header, but it also removes the date index, which is not what I want:
df3 = pd.read_csv(BytesIO(df1.to_csv(index=False)))
df3
Out[84]:
A B
0 2 test1a test1b
1 1 test2a test2b
Summing up, I wonder how one removes the index column header without removing the rest of the index column when applying pd.read_csv and to_csv. I want the df that is loaded using pd.read_csv to be identical to df1 in the example above.
