Put Numbering before Every Matched String Pattern

Question

I need to replace this:

01:05:01:11 --> 01:05:04:07,so you may continue to support us,|bring us health,
$Italic = True
01:05:04:15 --> 01:05:07:09,well-being,
$Italic = False
01:05:07:21 --> 01:05:13:01,and help us to be one big family|and continue working as a team.

To become essentially this:

1
01:05:01:11 --> 01:05:04:07,so you may continue to support us,|bring us health,
$Italic = True
2
01:05:04:15 --> 01:05:07:09,well-being,
$Italic = False
3
01:05:07:21 --> 01:05:13:01,and help us to be one big family|and continue working as a team.

EDIT_1: Which means that I need to match:

' --> '

And count its occurrences.

EDIT_2: So, for example, I need to match only the lines which contain:

01:05:04:15 --> 01:05:07:09,

And before each such line I need the number of the occurrence of the aforementioned example to be inserted into the file.

I've come up with this short Shell script which utilizes 'sed' command but it takes ages to process a bit bigger file (more than 60 lines, for example).

# Define the number of the special chars - so you can calculate the number of the subtitle lines
special_chars_no="$(grep -o ' --> ' Output_File | wc -l)"

# Add numbering before every subtitle line
for ((i=1;i<=${special_chars_no};i++)) ;
do
sed -i '/\([0-9][0-9]\):\([0-9][0-9]\):\([0-9][0-9]\):\([0-9][0-9]\) -->/{:1 ; /\(.*\([0-9][0-9]\):\([0-9][0-9]\):\([0-9][0-9]\):\([0-9][0-9]\) -->\)\{'"${i}"'\}/!{N;b1} ; s/\([0-9][0-9]\):\([0-9][0-9]\):\([0-9][0-9]\):\([0-9][0-9]\) -->/'"${i}"'\n\1:\2:\3:\4 -->/'"${i}"' ; :2 ; n ; $!b2}' Output_File  
done

Can we make it usable (much faster)?

Answer 1

$ awk '/-->/{print ++cnt} 1' file
1
01:05:01:11 --> 01:05:04:07,so you may continue to support us,|bring us health,
$Italic = True
2
01:05:04:15 --> 01:05:07:09,well-being,
$Italic = False
3
01:05:07:21 --> 01:05:13:01,and help us to be one big family|and continue working as a team.

Answer 2

sed is not suitable when it comes to using arithmetic and using a shell loop to process text is not advisable

$ cat ip.txt 
01:05:01:11 --> 01:05:04:07,so you may continue
$Italic = True
01:05:04:15 --> 01:05:07:09,well-being,
$Italic = False
01:05:07:21 --> 01:05:13:01,and help us to be

$ awk '/-->/{$0 = ++i RS $0} 1' ip.txt
1
01:05:01:11 --> 01:05:04:07,so you may continue
$Italic = True
2
01:05:04:15 --> 01:05:07:09,well-being,
$Italic = False
3
01:05:07:21 --> 01:05:13:01,and help us to be

• /-->/ if line matches this REGEXP
  • $0 = ++i RS $0 prefix input record with line number and separate them with value of RS which is newline by default
  • i variable will get 0 as default value in numeric context, ++i will give incremented value every time line matches the given REGEXP
• 1 idiomatic way to print contents of input record $0
• See also awk save modifications in place

You can also use perl

# use perl -i -pe for inplace editing
perl -pe 's/^/++$i . "\n"/e if /-->/' ip.txt
# or, borrowing Ed Morton's simplicity
perl -lpe 'print ++$i if /-->/' ip.txt

Answer 3

This might work for you (GNU sed):

sed -r '/-->/{x;:a;s/9(_*)$/_\1/;ta;s/^_*$/0&/;s/$/\n0123456789/;s/([^_])(_*)\n.*\1(.).*/\3\2/;y/_/0/;G;p;s/\n.*//;x;d}' file

On encountering the string -->, swap to the hold space (HS) and replace any trailing 9's with _'s. Add a 0 if this is the first time or all the characters are all _'s. Increment the last number and then replace all _'s by 0's. Append the pattern space (PS) and print the counter and the current line. Remove the current line, leaving the counter primed for the next match and return to the PS. Finally delete the PS. For lines that do not match, print as normal.

Answer 4

Your question is not that clear, seeing your ecpected output, following awk may help you in same.(I have old awk so adding re-interval in recent awk it could be removed.) I am assuming you want to look a specific string on a line and print the line number of it.

awk --re-interval '/[0-9]{2}:[0-9]{2}:[0-9]{2}:[0-9]{2}/{print FNR ORS $0}'  Input_file

In case you want to add count before a line then change ORS to OFS in above code.

In case you need to save output in your Input_file itself then following may help you in same too.

awk --re-interval '/[0-9]{2}:[0-9]{2}:[0-9]{2}:[0-9]{2}/{print FNR ORS $0}'  Input_file > temp_file && mv temp_file  Input_file

EDIT: If you want to only print line number before each line then following may help you.

awk '{print FNR ORS $0 ORS}'  Input_file