I have about a million tables stored as .txt files in a directory. The content of the files look like this:

Code:

ages-eligible-for-study:
genders-eligible-for-study:
accepts-healthy-volunteers:
CONTINUE

18-months-to-36-months
both
yes
CONTINUE
DONE

Other examples,

Code:

CONTINUE
non-electric-grass-pitch
electric-grass-pitch
towan-two-bedroom-holiday-home
crantock-two-bedroom-apartment
CONTINUE
10-x
10-x
6-x
4-x
CONTINUE
DONE

My aim is to change the format of text in each file to a tabular form. The output that I expect is:

Code:

ages-eligible-for-study: 18-months-to-36-months
genders-eligible-for-study: both
accepts-healthy-volunteers:yes

Code:

non-electric-grass-pitch 10-x
electric-grass-pitch 10-x
towan-two-bedroom-holiday-home 6-x
crantock-two-bedroom-apartment 4-x

Note that above is an example with just two columns, there can be tables more than two columns, like the one below:

Code:

size
waist-(on-half)
hip-(on-half)
inseam
CONTINUE
alt
14"
24-3/4"
34"
CONTINUE
axlt
15"
26-3/4"
34-1/2"
CONTINUE
a2xlt
16"
28-3/4"
35"
CONTINUE
a3xlt
17"
30-3/4"
35-1/2"
CONTINUE
DONE

I have written a bash script, which does the following in a for loop for each file:
1. Count number of times CONTINUE occurs in a file:

Code:

num_char=`grep -w 'CONTINUE' filename | wc -l`

2. Save text above CONTINUE in separate temp files using this command, which does not seem to work:

Code:

awk '/CONTINUE/{x="temp_file"++i;next}{print > x;}' $file

3. Delete DONE from the "temp_file"'s using this command:

Code:

sed -i '/DONE/d' ./$temp_file

4. Paste the contents in another file but in a same order as the columns occur in the main file, which I am finding it difficult to do too. I think I need another for loop here:

Code:

paste -d ' ' temp_file1 temp_fil2 > output_file

5. Delete all blank lines from output_file

Code:

awk 'NF' output_file > tmp && mv tmp ./output_file

The above process, as I suspect, will be too slow too on millions of files. I also wrote a Python program, but that is also not giving the desired output. I am using Linux and BASH.