With pandas.read_csv, I am reading a CSV which includes null entries, e.g.,
col1,col2,col3
2,3,4
8,,9
read_csv interprets the null entry as a float nan (since there is no int nan), so the column is being read as float64. If I try to force the column to be read as an integer, I get an error:
pd.read_csv(FILE, dtype = {'col2': int})
=> ValueError: Integer column has NA values in column 1
Should I read the data as float or string and convert afterwards? Or is there some way to specify that a null entry in an int column should be interpreted as zero?
Also, is there some way to turn on more detailed error reporting? I am new to pandas and it would have been helpful to know that the error was in row 1 (obviously my real data was much bigger).
question from:https://stackoverflow.com/questions/65865838/how-do-i-get-pandas-read-csv-to-interpret-a-null-entry-as-integer-0
