CODEWITHSUNDEEP
pythonpython-3.xpandasdataframe
Sven
Sven

Reputation: 101

Why can i not rename the columns?

df= get_returns("SP500")
df1=df["Close"]

df2 = get_returns("NASDAQ")
df2 =df["Close"]

asset = pd.concat([df1,df2],axis=1)
asset.rename({"Close": "Close SP00","Close": "Close NASDAQ"})
print(asset)

i also tried it with

asset.rename(columns={"Close": ""Chg_Close SP00","Close": "%Chfg_Close NASDAQ"})

and with

df1.columns("")
df2.columns("")

what is suspicious is that if i print just one of the datasets than there is no column name. But i do not why, because bfore python knew what was df2 = get_returns("NASDAQ") & df2 =df["Close"]

Date
2021-03-12         NaN
2021-03-11    0.016169
2021-03-10   -0.044696
2021-03-09   -0.014814
2021-03-08   -0.025117
                ...   
2020-03-18   -0.028696
2020-03-17    0.221267
2020-03-16    0.016338
2020-03-13    0.263508
2020-03-12   -0.070825

if i print the last row

print(asset)

then Column is shown up each but the change of renaming both didnt work.

               Close     Close
Date                          
2021-03-12       NaN       NaN
2021-03-11  0.016169  0.016169
2021-03-10 -0.044696 -0.044696
2021-03-09 -0.014814 -0.014814
2021-03-08 -0.025117 -0.025117
...              ...       ...

Upvotes: 0

Views: 67

Answers (2)

perl
perl

Reputation: 9941

what is suspicious is that if i print just one of the datasets than there is no column name

What happens is that your df1 and df2 are pd.Series and not pd.DataFrame, so you don't have a column name for them (they do have a name though, if you print them you should see Name: Close at the bottom of the output).

Probably the best option is to simply create a new DataFrame and specify the column names that you want:

asset = pd.DataFrame({
    'Close SP500': df1,
    'Close NASDAQ': df2,
})

P.S. And if you want to rename columns in a DataFrame, you need to assign it back to the DataFrame (df = df.replace(...)) or add inplace=True parameter (df.replace(..., inplace=True), otherwise your original DataFrame is not gonna be updated. If you have duplicate column names, however, it's best to simply set them with df.columns = [...], as suggested in the other answer

Upvotes: 1

Sven
Sven

Reputation: 101

def get_returns(file):
    dfcolumns = pd.read_csv(file + ".csv",nrows=1)

    return pd.read_csv(file + ".csv", index_col = 0, parse_dates = True,dtype={"Open": float,
                                                                              "High":float,
                                                                               "Low":float,
                                                                               "Close":float},usecols = list(range(len(dfcolumns.columns)))).pct_change()
#example

this is the get_returns function. I had to apply

dfcolumns = pd.read_csv(file + ".csv",nrows=1)

because the program always responded with these errors

TypeError: unsupported operand type(s) for /: 'str' and 'float'
result[mask] = op(xrav[mask], yrav[mask])

and then b)

 IndexError: list index out of range

Upvotes: 0

Related Questions