Why isnt df["ColumA"] = df_2["ColumnB"] working as expected?

Question

given two DataFrames:

#df1
value         Symbol       Date   Type        Value
1109914    AUD 2016-01-29  value   64587.9121    
1109949    CAD 2016-01-29  value   65357.1429
1109970    CHF 2016-01-29  value  111881.8681
1110012     DX 2016-01-29  value   91256.4103

#df2
volas         Symbol       Date  Type     Value
1109915    AUD 2016-01-29  vola  518.5091
1109950    CAD 2016-01-29  vola  360.9474
1109971    CHF 2016-01-29  vola  767.1880
1110013     DX 2016-01-29  vola  495.8913
1110041    EUR 2016-01-29  vola  876.6861
1110062    GBP 2016-01-29  vola  462.6425

If I want to add a columnd to df1 from df2, naturally, I'd try this:

df1["volas"] = df2["Value"]

However, this returns a Series with nan values (and hence fills the new column with such):

1109921   NaN
1110089   NaN
1110138   NaN
1110159   NaN
1110257   NaN
1110450   NaN
1110618   NaN
1110667   NaN
1110688   NaN

However, if I do this:

market_values['Volas'] = [i for i in market_volas.Value]

The values are assigned as expected.

Why? I simply cannot see where the problem is, especially because my initial attempt is suggested here.

Answer 1

There is problem that index in df1 is different as index in df2.

If length od indexes is equal, you can use values:

df1["volas"] = df2["Value"].values
print (df1)
     value Symbol        Date   Type        Value     volas
0  1109914    AUD  2016-01-29  value   64587.9121  518.5091
1  1109949    CAD  2016-01-29  value   65357.1429  360.9474
2  1109970    CHF  2016-01-29  value  111881.8681  767.1880
3  1110012     DX  2016-01-29  value   91256.4103  495.8913

Answer 2

Think of

df1["volas"] = df2["Value"]

as doing a merge on the index of the two dataframes. What you're probably after is an actual merge based on some matching columns, e.g.

df1.merge(df2, on=["Symbol","Date"])

which will add the columns from df2 based on matches with the other columns, or, if they are of the same length

df1["volas"] = df2.Value.values

which is basically your loop solution.