如何将字符串写入CSV文件

Question

I am trying to create a csv file with a single column of file paths. I need a second column filled with ones.

The result I want to get is as follows;

./1/a_1.csv, 1
./1/a_2.csv, 1
./1/a_3.csv, 1

The code I tried is this;

import numpy as np

data=np.chararray((650,2), itemsize=20)

for i in range(1, 650):
    data[0][0]="./1/a_" + str(i) + ".csv"

np.savetxt("C:\\a.csv", data, delimiter = ",")

However, it doesn't seem to work. Could you help?

Answer 1

You don't need Numpy for this. Just do something like

with open('a.csv', 'w') as outf:
    for i in range(650):
        print('./1/a_%s.csv, 1' % (i + 1), file=outf)

and you're golden.

Answer 2

If you don't care about how the file is created (as long as it works), go for AKX's solution.

If you want to do it with Numpy, you need to get your dtype straight. np.savetxt has a default formatter which has trouble understanding the array you gave it.

This is how I managed to get your desired output:

data = np.empty((650,2),dtype="S20") # empty array of the correct datatype 

data[:,1] = '1' # set column 1 to '1' (as you noted)

for i in range(650):
    data[i][0] = "./1/a_" + str(i) + ".csv" # put your filenames into first column

np.savetxt(r"C:\a.csv", data, delimiter = ",", fmt="%s") # save it using %s formatter

Answer 3

In each iteration of your loop, you only ever write to data[0][0] . You need to alter your loop to change the array position you're setting as you loop over the range.

Answer 4

In csv file the delimiter is ";" not ",".

import numpy as np

data=np.chararray((650,2), itemsize=20)

for i in range(1, 650):
    data[0][0]="./1/a_" + str(i) + ".csv"

np.savetxt("C:\\a.csv", data, delimiter = ";")

Answer 5

I would use Dataframes from the Pandas library. A similar approach to chitown88's answer, but more compact.

import pandas as pd

# create a list with paths to files named from 1 to 256
file_paths = ["./1/a_{}.csv".format(i) for i in range(1,257)]

# add the list to a Pandas Dataframe, the column name can be omitted
df = pd.DataFrame(data=file_paths, columns=['Paths'])

# add a column of 1's to the DataFrame
df['1'] = 1

# write the DataFrame without the indices and column names to a csv file using a specific separation character  
df.to_csv('C:\\a.csv', sep=',', index=False, header=False)

Answer 6

I personally like to store all my data into a dataframe and then at he very end, save it to a csv file, as opposed to writing after each iteration.

so I initialize a blank dataframe called results

then iterates through your range, creating your file_path string. Then I create that 1 row with the current file_name string and your 2nd column. That gets appended to the results dataframe, then moves on to the next item in your list/range. Continues that process until it's done, then saves the final results dataframe to file

import pandas as pd

# initialize an empty dataframe 
results = pd.DataFrame()

# iterates each element in the range list to form a row that is appended to the results dataframe
for i in range(1, 650):
    file_path="./1/a_" + str(i) + ".csv"
    temp_df = pd.DataFrame([[file_path, 1]], columns = ['file_path', 'col_2'])
    results = results.append(temp_df.reset_index(drop = True))

# after iterating through all the elements, saves completed results dataframe to file
results.to_csv("a.csv", index=False))