convert non digits in one column to nan in pandas-CodePudding

My pandas data frame is as below. IN the "price" column, I want to convert non digit values to nan. Thanks.

Input:

ID, name, qty, price 
1, pen, 4, 55
2, book, 5, ai
3, PENCIL, 5, 100
4, book2, 10, si

output:

ID, name, qty, price 
1, pen, 4, 55
2, book, 5, nan
3, PENCIL, 5, 100
4, book2, 10, nan

CodePudding user response：

Use pd.to_numeric:

df['price'] = pd.to_numeric(df['price'], errors='coerce')

If they're strings, use mask and str.isnumeric:

df['price'] = df['price'].mask(~df['price'].str.isnumeric())

mask automatically converts data that match the condition to NaN.

Or use str.isaplha:

df['price'] = df['price'].mask(df['price'].str.isalpha())

CodePudding user response：

You could use str.replace here:

df["price"] = df["price"].str.replace(r'^.*\D.*$', np.NaN)

The regex pattern ^.*\D.*$ will match any price entry having at least one non digit character, and will selectively replace with nan.