I have the following dataframe:

symbol  Open  close
SPY  34,2  33,2
AMZN 30.2  10,2
.................

and I want to create a function that will convert each character in a string into ASCII. This function will than be applied on symbol field and its value will be placed in a new field (called 'id'). The final dataframe will look:

symbol  Open  close id
SPY  34,2  33,2  838089
AMZN 30.2  10,2  65779078

this is what I have done

def symbolid(x):
    strAscii = ''
    for i in range (len(x)):
        strAscii =  strAscii   str(ord(x[i]))
        print(x)
    return strAscii

df['id'] = df.apply(lambda x: symbolid(df['symbol']), axis=1)

I get TypeError: ord() expected a character, but string of length 3 found

any help would be appreciate/ Using python 3.6

CodePudding user response：

In your lambda function you use whole column as argument, that's source of error, you can only use symbol column.

    import pandas as pd

data = {
    "symbol": ["SPY","AMZN"],
     "Open": [34.2 , 33.2],
     "close":[30.2  ,10.2]
}
df = pd.DataFrame(data)
def symbolid(x):
    strAscii = ''
    for i in range (len(x)):
        strAscii =  strAscii   str(ord(x[i]))
        print(x)
    return strAscii
df['id'] = df['symbol'].apply(lambda x: symbolid(x))
print(df)

CodePudding user response：

You were not that far. But in your lambda function you still refere df while you should use x:

df['id'] = df.apply(lambda x: symbolid(x['symbol']), axis=1)

shows

SPY
SPY
SPY
AMZN
AMZN
AMZN
AMZN

and gives as expected

df
  symbol  Open close        id
0    SPY  34,2  33,2    838089
1   AMZN  30.2  10,2  65779078

But your code is not very efficient because:

1. in your function you add to a string when you could join an iterator
2. you convert each row into a Series when you could simply transform one column

In the end,

df['symbol'].transform(lambda x: ''.join(str(ord(i)) for i in x))

Also gives the expect dataframe