linux - 将具有不同行条目的 txt 文件转换为.csv？

Question

I have used exiftool to make.txt files with the EXIF data of some images and videos I am working with. I will also eventually need to create.csv manifests for these files, and I know there are simple ways to convert.txt files to.csv files, but the instructions I've found have described how to do so if the.txt file has the information to go in different columns within the same line, while mine are on different lines. Is there a way to do this conversion with.txt files that are organized differently?

For example, I have seen instructions for how to convert something like this

filename   date       size
abc.JPG    1/1/2001   1MB
def.JPG    1/1/2001   1MB
hij.JPG    1/1/2001   1MB

to

filename,date,size
abc.JPG,1/1/2001,1MB
def.JPG,1/1/2001,1MB
hij.JPG,1/1/2001,1MB

The.txt files I have, on the other hand, are formatted like this:

========abc.JPG
File Name            abc.JPG
Date/Time Original   2001:01:01 1:00:00
Size                 1 MB
========def.JPG
File Name            def.JPG
Date/Time Original   2001:01:01 1:01:00
Size                 1 MB
========hij.JPG
File Name            hij.JPG
Date/Time Original   2001:01:01 1:02:00
Size                 1 MB

but I still need an output like

filename,date,size
abc.JPG,2001:01:01 1:00:00,1 MB
def.JPG,2001:01:01 1:01:00,1 MB
hij.JPG,2001:01:01 1:02:00,1 MB

Answer 1

Using sed and Miller , you can run

<input.txt sed -r 's/==.+//g;s/([a-zA-Z]) ([a-zA-Z])/\1-\2/' | mlr --x2c label filename,date,size >output.csv

to have

filename,date,size
abc.JPG,2001:01:01 1:00:00,1 MB
def.JPG,2001:01:01 1:01:00,1 MB
hij.JPG,2001:01:01 1:02:00,1 MB