Get string from file - linux

Question

In my Packages file i have multiple packages. I'm able to check the file if a string is inside, and if so, i would like to get the version of the file.
Package: depictiontest
Version: 1.0
Filename: ./debs/com.icr8zy.depictiontest.deb
Size: 810
Description: Do not install. Testing Depiction.
Name: Depiction Test
so the above is part of the many similar looking info of a package. Each time i detected if the package exists i would like to get the Version. is there any possible way?
btw, this is what i use to get check if the file exists.
if grep -q "$filename" /location/Packages; then
#file exists
#get file version <-- stuck here
else
#file does not exists
fi
EDIT:
Sorry but maybe i wasn't clear in explaining myself, I would already have the Name of the package and would like to extract the Version of that package only. I do not need a loop to get all the Names and Versions. Hope this clears it... :)

Answer 1

How do you extract the file name in the first place? Why not parse the whole file, then filter out nonexistent file names.
awk '/^Package:/{p=$2}
/^Version:/{v=$2}
/^Filename:/{f=$2}
/^$/{print p, v, f}' Packages |
while read p v f; do
test -e "$f" || continue
echo "$p $v"
done
This is not robust with e.g. file names with spaces, but Packages files don't have file names with spaces anyway. (Your example filename is nonstandard, though; let's assume it's no worse than this.)
You want to make sure there's an empty line at the end of Packages, or force it with { sed '$/^$/d' Packages; echo; } | awk ...
Edit: This assumes a fairly well-formed Packages file, with an empty line between records. If a record lacks one of these fields, the output will repeat the value from the previous record - that's nasty. If there are multiple adjacent empty lines, it will output the same package twice. Etc. If you want robust parsing, I'd switch to Perl or Python, or use a standard Debian tool (I'm sure there must be one).

Answer 2

With grep, you can pick a certain amount of lines before or after a keyword.
egrep -A1 "^Package: depictiontest" /path/to/file
would yield 1 additional line after the match.
egrep -B1 "^Filename: .*depictiontest.*" /path/to/file
would yield 1 additional line before the match.
egrep "^(Package|Version): " "^Package: depictiontest" /path/to/file
would lead to package and Version lines only, so would rely on them being in the correct order, to find out easily, which version belongs to which package.

Answer 3

If the order is same then you can parse the whole file and feed values in to an array -
awk -F": " '
/^Package/{p=$2;getline;v=$2;getline;f=$2;ary[p"\n"v"\n"f"\n"]}
END{for (x in ary) print x}' file
Test:
[jaypal:~/Temp] cat file
Package: depictiontest
Version: 1.0
Filename: ./debs/com.icr8zy.depictiontest.deb
Size: 810
Description: Do not install. Testing Depiction.
Name: Depiction2fdf Test
Package: depi444ctiontest
Version: 1.05
Filename: ./debs/coffm.icr8zy.depictiontest.deb
Size: 810
Description: Do not install. Testing Depiction.
Name: Depiction Test
Package: depic33tiontest
Version: 1.01
Filename: ./d3ebs/com.icr8zy.depictiontest.deb
Size: 810
Description: Do not install. Testing Depiction.
Name: Depiction Test
[jaypal:~/Temp] awk -F": " '/^Package/{p=$2;getline;v=$2;getline;f=$2;ary[p"\n"v"\n"f"\n"]}END{for (x in ary) print x}' file
depi444ctiontest
1.05
./debs/coffm.icr8zy.depictiontest.deb
depic33tiontest
1.01
./d3ebs/com.icr8zy.depictiontest.deb
depictiontest
1.0
./debs/com.icr8zy.depictiontest.deb

Answer 4

If "Version: ..." line is always exactly 1 line before the "Filename: ..." line, then you may try something like this:
line_number=$(grep -n "$filename" /location/Packages | head -1 | cut -d: -f1)
if (( $line_number > 0 )); then
#file exists
version=$(head -n $(( $line_number - 1 )) /location/Packages | tail -1 | cut -d' ' -f2)
else
#file doesn't exist
fi

Answer 5

The simplest awk implementation that I can think of:
$ awk -F':' -v package='depictiontest' '
$1 == "Package" {
trimmed_package_name = gensub(/^ */, "", "", $2)
found_package = (trimmed_package_name == package)
}
found_package && $1 == "Version" {
trimmed_version_number = gensub(/^ */, "", "", $2)
print trimmed_version_number
}
' Packages
1.0
This processes the file (Packages) line-by-line and sets the found_package flag if the line starts with 'Package' and the value after the field separator (-F), which is : (and any whitespace) is the value passed into the package variable (-v). Then, if the flag is set an we find a line beginning with a 'Version' field, we print the value after the field separator (trimming leading whitespace). If another 'Package' field is found and the name is not the one that we are looking for, the flag is reset and the subsequent version number will not be printed.