Using awk/sed to convert linefeed to csv, with some formatting

jaykup · 04-03-2009, 03:44 PM

I have a file that looks like the one below, only much larger:

Code:

2009-02-20
this is some text
145 sometimes numbers and text with spaces.


2009-02-07
this is some different text
276 sometimes numbers and text with spaces.


2007-12-18
thisisoneword
13 sometimes numbers and text with spaces.

I would like to change it like this:

Code:

2009-02-20, "this is some text", 145, "sometimes numbers and text with spaces."
2009-02-07, "this is some different text", 276, "sometimes numbers and text with spaces."
2007-12-18, "thisisoneword", 13, "sometimes numbers and text with spaces."

Top line, always a date in that format.
Middle line, text or numbers
bottom line, always an integer, a space, then random text. I want to seperate the first number from the rest with a comma.

The pattern and spacing is exactly like above.

Kenhelm · 04-03-2009, 05:18 PM

Using GNU sed

Code:

sed  '/./ {
N
N
s/\(.*\)\n\(.*\)\n\([0-9]\+\) \(.*\)/\1, "\2", \3, "\4"/
}
/^$/d' infile > outfile