Extract multiple values from fixed pattern string-CodePudding

I want to parse 3 pieces of information from the last bit of wget command output. For example:

2022-12-26 19:14:44 (13.7 Mb/s) - ‘somelibrary.min.js’ saved [1077022]

I was able to get the date/time since that is a fixed length. I am unable to extract the est. transfer speed (13/7) and file size (1077022) values.

STR="2022-12-26 19:14:44 (13.7 Mb/s) - ‘somelibrary.min.js’ saved [1077022]"
echo date/time is ${STR::19}

I imagine the remaining substring extractions will need to be done with the help of regular expressions, but I am unable to figure it out. Is there a viable path using only *nix utils like awk, sed, etc.?

I tried awk:

echo "(13.7 Mb/s)" | awk '$0 ~ /(.* Mb\/s)/ {print $1}'

But I am getting (13.7 instead of just the number.

CodePudding user response：

You could do this with bash's regular expression matching, using ( ) in the RE to capture the relevant parts and ${BASH_REMATCH[n]} to get them:

str="2022-12-26 19:14:44 (13.7 Mb/s) - ‘somelibrary.min.js’ saved [1077022]"

pattern='([-0-9]  [:0-9] ) \(([^)] )\) .*\[([0-9] )\]'
if [[ "$str" =~ $pattern ]]; then
    echo "date/time is ${BASH_REMATCH[1]}"
    echo "transfer speed is ${BASH_REMATCH[2]}"
    echo "file size is ${BASH_REMATCH[3]}"
else
    echo "The string is not in the expected format"
fi

BTW, I recommend using lower- or mixed-case variable names to avoid conflicts with the many all-caps names with special functions, and running your scripts through shellcheck.net to find common mistakes.

CodePudding user response：

This awk should work for you:

s="2022-12-26 19:14:44 (13.7 Mb/s) - ‘somelibrary.min.js’ saved [1077022]"
awk -F '[][()[:blank:]] ' '{
  printf "DateTime: %s %s, Speed: %s, Size: %s\n", $1, $2, $3, $(NF-1)
}' <<< "$s"

DateTime: 2022-12-26 19:14:44, Speed: 13.7, Size: 1077022

Breakdown:

-F '[][()[:blank:]] ' sets 1 of [ or ] or ( or ) or a whitespace as input field separator

CodePudding user response：

With your shown samples please try following awk code. Written and tested in GNU awk. Here is the Online Demo for used regex.

s="2022-12-26 19:14:44 (13.7 Mb/s) - ‘somelibrary.min.js’ saved [1077022]"

awk '
match($0,/([0-9]{4}-[0-9]{2}-[0-9]{2} [0-9]{2}:[0-9]{2}:[0-9]{2}) \(([^)]*).*\[([0-9] )/,arr){
  print "DateTime: "arr[1] ", Speed: " arr[2] ", Size: "arr[3]
}
' <<< "$s"

CodePudding user response：

Using Perl and named capture groups:

s="2022-12-26 19:14:44 (13.7 Mb/s) - ‘somelibrary.min.js’ saved [1077022]"

perl -nE '
    say join "\n",
    map { "$_: $ {$_}" }
    keys % 
    if m/
        (?<dateTime>\d{4}-\d{2}-\d{2}\s \d{2}:\d{2}:\d{2})\s \(
        (?<speed>\d \.\d ).*\[
        (?<size>\d )
    /x
' <<< "$s"

Output

speed: 13.7
dateTime: 2022-12-26 19:14:44
size: 1077022

Explanations

See regex101 description

CodePudding user response：

Using sed with 3 capture groups \1, \2 and \3 in the replacement:

sed -E 's/([^()] ) \(([^()]*)\).*\[([^][]*)]/date\/time: \1\nspeed: \2\nfile size: \3/' file

The pattern matches:

([^()] ) Capture group 1, match any character except for ()
$([^()]*)$ Between (...) capture any char except ( and ) in group 2
.* Match the rest of the line
\[([^][]*)] Between [...] Capture any char except [ and ] in group 3

See the capture group values here on regex101

Output

date/time: 2022-12-26 19:14:44
speed: 13.7 Mb/s
file size: 1077022