The original csv file data is like that:
06/04/2011,104.64,105.17
07/04/2011,104.98,105.51
08/04/2011,105.43,105.96
11/04/2011,104.47,104.99
How to either read the csv file into DataFrame and add multiple row index level, or add multiple row index into csv and import into DataFrame as following:
JAS
date bid ask
06/04/2011 104.64 105.17
07/04/2011 104.98 105.51
08/04/2011 105.43 105.96
11/04/2011 104.47 104.99
Read the CSV, setting the first (0th) columns as the index.
In [8]: df = pd.read_csv(StringIO("""06/04/2011,104.64,105.17
07/04/2011,104.98,105.51
08/04/2011,105.43,105.96
11/04/2011,104.47,104.99"""), index_col=0, header=None)
Create a new MultiIndex, and assign it to the columns.
In [11]: df.columns = pd.MultiIndex.from_tuples([('JAS', 'bid'), ('JAS', 'ask')])
Finally, name the index, and we have your desired result.
In [12]: df.index.name = 'date'
In [13]: df
Out[13]:
JAS
bid ask
date
06/04/2011 104.64 105.17
07/04/2011 104.98 105.51
08/04/2011 105.43 105.96
11/04/2011 104.47 104.99
Short answer:
df = pd.read_csv('file.csv', parse_dates=True, index_col=0, header=None).rename_axis(
index='date').rename(columns={1: 'bid', 2: 'ask'}).reindex(
columns=pd.MultiIndex.from_product([['JAS'], ['bid', 'ask']]), level=1)
Out[1]:
JAS
bid ask
date
2011-06-04 104.64 105.17
2011-07-04 104.98 105.51
2011-08-04 105.43 105.96
2011-11-04 104.47 104.99
