Get 2 texts from this line:
[23/08/2019 09:30:00 351] [DEBUG] [Hilo:Thread-17905] [Clase:co.com.colombiamovil.activator.db.PoolFacade: getSiebelDirectConnection (81)] [Conexi???n establecida 7682 siebelDB: TMAC70580369126846, duraci???n: 90]
Output:
23/08/2019 09:30:00 | duraci???n: 90
or
23/08/2019 09:30:00 | 90
With this I take the duration:
grep "duraci???n: " ussdGw.log | cut -d" " -f16 | tr -d "]"
but I can't get the date out
Help me
You could just use awk for this:
$ awk 'gsub(/[][]/,""){print $1,$2,"|",$NF}' file
23/08/2019 09:30:00 | 90
The gsub(/[][]/,"") will replace all occurrences of either [ or ] (that's a character class: [ ] consisting of two characters ] and [, so becomes [ ][ ]) on the current line and then, if that was successfull, you print the 1st field (the date), the second field (the time) and the last (NF) field, the duration.
I guess you want to parse a logfile with lots of lines? Making use of substring-manipulation:
$ cat get2texts.sh:
#!/bin/bash
logfile="$1"
IFS=']'
while read -r one two three four five junk; do
one="${one#[}"
one="${one% *}"
five="${five/*,/ | }"
echo "$one $five"
done < "$logfile" \
| tr -s ' ' ' '
Result:
$ get2texts.sh testfile
23/08/2019 09:30:00 | duraci???n: 90
If you set FS sensibly, you might get away with e.g.:
awk '{ print $2, $3, "| " $(NF-1) }' FS='[] []+' infile
Output:
23/08/2019 09:30:00 | 90
To get the date and duration in lines containing duraci???n: with sed:
$ sed -n 's/^\[\([^ ]* [^ ]*\).*\(duraci???n:.*\)\]$/\1 | \2/p' logfile
23/08/2019 09:30:00 | duraci???n: 90
or
$ sed -n 's/^\[\([^ ]* [^ ]*\).*duraci???n: \(.*\)\]$/\1 | \2/p' logfile
23/08/2019 09:30:00 | 90
Explanation of the first command:
sed -n be silent (only matched parts are printed with the p flag)'s/ substitute^\[ match [ at the start of the line\( start first group[^ ]* [^ ]* match date and time (any no space characters followed by space followed by any no space characters)\) end first group.* match any characters\( start second groupduraci???n:.* match duraci???n:.* followed by any characters\) end second group\]$ match ] at the end of the line/\1 | \2/ replace with first group, | and the second groupp' print matches
awk '{print $1,$2,$NF}' filename| sed "s/\[//g"|sed "s/\]//g"| awk '{$NF=" | "$NF;print $0}'
output
23/08/2019 09:30:00 | 90
