Search code examples
pythonnumpycsv

How to read numbers from csv file to calculate the mean of a column and skip empty strings?

I have a csv file and I have to compute the mean for some of the columns. That's how I did:

file = csv.reader(open('tab.csv','r'))
n = []
for row in file:
    n.append(row[8])

So I have a list of string: n = ['', '', '1.58', ...]. How can I convert these to float? I tried with:

n_values = np.array(n)
n_values[n == ''] = '0'
values = n_values.astype(np.float)
np.mean(values)

But the mean is not correct because I should skip the empty strings not counting.

Solution

  • Just cast as you append:

     n.append(float(row[8]))

    If there are empty strings catch those before appending.

    try:
    n.append(float(row[8]))
except ValueError:
   continue

    Or you might want to try pandas, in particular pandas.read_csv:

    import pandas as pd

df = pd.read_csv("in.csv")
print(df["col_name"].mean())