value error while converting object to int-CodePudding

my quantity dtype is object (from csv file) and I am trying to covert to int using below code:

df[x1] = df[x1].astype(str).astype(int)

It throws error as below:

ValueError: invalid literal for int() with base 10: '1,000.000'

Can anyone help me in this please?

CodePudding user response：

You need to remove ',' and '.'. You can use str.replace method to remove the comma and then cast the data to type float then to type int.

df[x1] = df[x1].str.replace(',','').astype(float).astype(int)

For example, for a Series such as

srs = pd.Series(['1,000.00','1'])

if you cast it to dtype int

srs.astype(int)

you get

ValueError: invalid literal for int() with base 10: '1,000.00'

Then if you remove the comma str.replace method and cast to dtype int

srs = srs.str.replace(',','')
srs.astype(int)

you get

ValueError: invalid literal for int() with base 10: '1000.00'

So you cast it to dtype float and to dtype int,

srs = srs.str.replace(',','').astype(float).astype(int)

you get the expected outcome:

0    1000
1       1
dtype: int32

CodePudding user response：

Actually int() function expects an integer string or a float, but not a float string. If a float string is given you need to convert it to float first then to int as:

int(float(userGuess))

So you have to convert it to float first and then to int.