我有一个如下所示的 csv：
someFile.csv
Header1 Header2 Header3
aa aaa a
bbbb bbbbbb aa
我想计算每列的平均字符串长度并创建结果的 csv。这就是示例中的样子：
results.csv
Header1 Header2 Header3
3 4.5 1.5
我一直在尝试在 Python 中使用 csv 库，但没有成功。有没有一种简单的方法可以做到这一点？
【问题讨论】：
标签： python python-3.x csv
你可以试试熊猫。如果您没有安装 pandas，请通过pip install pandas 安装 pandas。
import pandas as pd
# df = pd.read_csv('my_csv.csv')
df = pd.DataFrame([['aa', 'aaa', 'a'], ['bbbb', 'bbbbbb', 'aa']],
columns=['Header1', 'Header2', 'Header3'])
result = pd.DataFrame([[]])
for col in df:
result[col] = df[col].apply(len).mean()
result.to_csv('result.csv')
希望这会有所帮助！
【讨论】：
您可以zip 行和map 列到len 并使用statistics.mean 计算平均值：
import csv
from statistics import mean
with open('someFile.csv', 'r', newline='') as f, open('results.csv', 'w', newline='') as output:
reader = csv.reader(f, delimiter=' ', skipinitialspace=True)
headers = next(reader)
writer = csv.writer(output, delimiter = ' ')
writer.writerow(headers)
writer.writerow([mean(map(len, col)) for col in zip(*reader)])
【讨论】：
这是一个简单的代码。我提供了两个块，如果数据帧中没有空值并且存在空值。
import pandas as pd
#df = pd.DataFrame([['aa','aaa','a'],['bbbb','bbbbbb','aa']],columns=['Header1','Header2','Header3'])
df = pd.read_csv('file.csv')
#if No Null
No_of_Row=df.shape[0]
for Col in df.sum():
print(len(Col)/No_of_Row)
#if Null are there
for Col,Header in zip(df.sum(),df.columns):
print(len(Col)/df[[Header]].dropna().shape[0])
【讨论】：
这不是最好的方法。还有其他方法可以快速做到这一点。但是，我确实认为这是一个相当直接且易于理解的示例，并且非常仓促地组合在一起。我在你的例子中使用了这个，它有效。
import csv
# replace "yourusername" with your PC user name
input_file = 'C:/Users/yourusername/Desktop/someFile.csv'
output_file = 'C:/Users/yourusername/Desktop/output.csv'
csv_file = open(input_file, newline='') # opening csv file
info = list(csv.reader(csv_file)) # convert data in csv file to array/list
csv_file.close()
length = len(info[0]) # if you ever add more headers, this will account for it
avg_container = [0 for i in range(length)] # creates empty array with zeros for each header
n = len(info[1:]) # for dividing by n to get average
# adding the lengths of all the items to one sum for each "column"
for k in info[1:]:
for n,i in enumerate(k):
avg_container[n] += len(i)
# diviving all sums by n
for i in range(len(avg_container)):
avg_container[i] = avg_container[i]/n
# combine header and average array into one item to write to csv
avg_output = []
avg_output.extend((info[0],avg_container))
print(avg_output) # just for you to see for yourself
# outputting the new file
output_csv = open(output_file, 'w', newline='') # creates an instance of the file
csv_writer = csv.writer(output_csv) # creates an "Writer" to write to the csv
csv_writer.writerows(avg_output) # outputs the avg_output variable to the csv file
output_csv.close() # finished
参考文献
How to import a csv-file into a data array?
【讨论】：