I have a CSV dataset with 2 columns that looks like the following:

Data columns (total 2 columns):

dtypes: datetime64ns, object(1)

When trying to pass this through a timeseries model I get the following error:

ftse_open = TimeSeries.from_dataframe(ftse_open, time_col='Date', value_cols='Open')

ValueError: could not convert string to float: '7,541.85'

Then I try a different route using the following code:

ftse_open["Open"] = ftse_open["Open"].astype('Int64')

Yielding:

TypeError: object cannot be converted to an IntegerDtype

I have tried more code to resolve but I'm not sure why there seems to be no solution that I can find.

(There are no NAs in the dataset - I have checked).

Any help is appreciated, thank you.

CodePudding user response：

Based on comments, you can try:

df["Open"] = df["Open"].str.replace(",", "").astype(float)
print(df)

Prints:

      Date     Open
0  25/2/21  7541.85
1  26/2/21  7562.32
2  27/2/21  7521.65
3  28/2/21  7509.14

df used:

      Date      Open
0  25/2/21  7,541.85
1  26/2/21  7,562.32
2  27/2/21  7,521.65
3  28/2/21  7,509.14