How to handle non-numeric entries in an integer valued column

Question

I have a dataframe df and one of the columns count is contains strings. These strings are mostly convertable to integers (e.g. 0006) which is what I will do with them. However some of the entries in count are blank strings of spaces. How can I

• Drop all the rows where the count value is a blank string.
• Substitute all the blank values in that column with some numeric value of my choice.

The dataframe is very large if there are particularly efficient ways of doing this.

Answer 1

It seems that you want two different things. But first, convert column to numeric and coerce errors:

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

To drop rows (use subset to avoid dropping NaN from other columns):

df.dropna(subset=['count'])

To replace with default value:

df['count'] = df['count'].fillna(default_value)