write pandas Dataframe row into text file

Vishvajit Jambuti 0
Viewed 57

I have data as shown in the picture, please check the dataframe here:
enter image description here

I want to convert that data amsterdam_01136.txt, amsterdam_01239.txt ... so on

which should look like this (yolo object detection annotation format)

0 0.982552 0.597656 0.034896 0.121094

0 0.742448 0.613770 0.028646 0.059570

2 0.418750 0.540527 0.038542 0.200195

I tried using for loop as given here (3rd solution) link but it only adds the last row of the file name. please give me a suggestion

text-files string dataframe pandas python
3 Answers
df.to_csv(r'destination_path',sep=',',columns=[list of columns],index=False)
zeeshan123960

Assuming your dataframe is this:

import pandas as pd
df = pd.DataFrame({'filename': ['amsterdam_01136.png', 'amsterdam_01136.png', 'amsterdam_01136.png'],
                   'class':    [0,0,0], 
                   'xcenter':  [0.982,0.742,0.434], 
                   'ycenter':  [0.597,0.613,0.574], 
                   'width':    [0.034,0.028,0.036],
                   'height':   [0.121,0.059,0.096]  })
print(df)

              filename  class  xcenter  ycenter  width  height
0  amsterdam_01136.png      0    0.982    0.597  0.034   0.121
1  amsterdam_01136.png      0    0.742    0.613  0.028   0.059
2  amsterdam_01136.png      0    0.434    0.574  0.036   0.096

What would you like to do with repeated filenames for the first 11 rows?

for col in df.columns:
    df[col] = df[col].astype('str')
for i in range(len(df)):
    filename = df.iloc[i, 0].rsplit('.', 1)[0] + '.txt'
    print(filename)
    line = ' '.join(df.iloc[i, 1:].tolist())
    print(line)
    with open(filename, 'w') as f:
        f.write(line)

amsterdam_01136.txt
0 0.982 0.597 0.034 0.121
amsterdam_01136.txt
0 0.742 0.613 0.028 0.059
amsterdam_01136.txt
0 0.434 0.574 0.036 0.096
perpetualstudent0

Thank You guys for help, I wanted the answer like this if anyone needs it

df = above given df    
f_names = df['filename'].unique()
f_names = f_names.tolist()
for f_name in f_names:
    new_df = df[df['filename']==f_name]
    save_df = new_df[['class', 'xcenter', 'ycenter', 'width', 'hight']] 
    fname_new  = f_name.split(".")[0]
    np.savetxt(f"/data/{fname_new}.txt", save_df, delimiter=" ", fmt='%s')
Vishvajit Jambuti0
Related