Updating dataframe by row but not updating

Question

I am iterating through a dataframe and I am trying to add values to a specific column for each row but when I print the resulting data frame the values are not present

#add two new blank columns to the dayData dataframe 
dayData["myValue1"]=""
dayData["myValue2"]=""

#iterate over the dataframe
for idxDay, row in dayData.iterrows():
        do something.....
        #interate again through the dataframe
        for idxRange, row1 in dayData.iterrows():
            do something else....
            calculate value1
            calculate value2 

        #write the result for value1 and value2 to the dayData dataframe         


        row["myValue1"]=value1
        row["myValue2"]=value2
        print(dayData)

The values for value1 and value 2 are correct and even when I hard code value1 = 1 and value2 = 2 when I print dayData the columns after columns myValue1 and myValue2 should be updated they contain no data.

the resulting dayData dataframe should look like

         vwap        last       volume     ratio myLong myShort  
0  301.071871  301.221525   43133218.0  1.000497   1       2       
1  215.545413  213.791400  349730738.0  0.991862   3       3

but instead I just get:

         vwap        last       volume     ratio myLong myShort  
0  301.071871  301.221525   43133218.0  1.000497                 
1  215.545413  213.791400  349730738.0  0.991862          

Answer 1

Instead of assigning to the row, assign back to the dataframe with index. The row is a copy of the row and changes to it won't persist in the parent dataframe.

Instead of:

        row["myValue1"]=value1
        row["myValue2"]=value2

Do:

        dayData.loc[idxDay, "myValue1"]=value1
        dayData.loc[idxDay, "myValue2"]=value2

Further example:

df = pd.DataFrame([1], ['a'], ['A'])

print df

   A
a  1

for i, r in df.iterrows():
    r.loc['B'] = 2

print df

   A
a  1


for i, r in df.iterrows():
    df.loc[i, 'B'] = 2

print df

   A    B
a  1  2.0

Clearly shows that assigning to row does not work. Assigning to dataframe does.