Remove string based on start and end pattern and remove newline in the process-CodePudding

I have a file that contains output of some commands - unfortunately some of them are mangled by console error:

path="/a/b/c" fxid="05200f4d" size="12550144" nrep="2" checksumtype="adler" checksum="045a6aa400000000000000000000000000000000"
nrep="01" fsid="132" host="cluster.local:1095" fstpath="/data/00019507/3dcd7e00" size="4574568" statsize="45745error: unable to retrieve file meta data from cluster.local:1095 [ status=down ]
nrep="00" fsid="37" host="cluster.local:1095" fstpath="/data/000021ca/0527e888" size="12550144" statsize="12550144" checksum="bb2a2ea700000000000000000000000000000000" diskchecksum="bb2a2ea700000000000000000000000000000000" error_label="nerror: unable to retrieve file meta data from cluster.local:1095 [ status=(null) ]
one"
path="/a/b/b98d6d3a-5c77-4223-9601-9294c73e00f9.bin" fxid="05200f4d" size="12550144" nrep="2" checksumtype="adler" checksum="045a6aa400000000000000000000000000000000"
nrep="01" fsid="36" host="cluster.local:1095" fstpath="/data/00002196/05200f4d" size="12550144" statsize="12550144" checksum="045a6aa400000000000000000000000000000000" diskchecksum="045a6aa400000000000000000000000000000000" error_label="nonerror: unable to retrieve file meta data from cluster.local:1095 [ status=(down) ]
e"
path="/a/b/c/.mb6589013703229118680.txt" fxid="0524071a" size="0" nrep="2" checksumtype="adler" checksum="0000000100000000000000000000000000000000"
nrep="00" fsid="196" host="cluster.local:1095" fstpath="/dataerror: unable to retrieve file meta data from cluster.local:1095 [ status=(null) ]
/000021b0/0524071a" size="0" statsize="0" checksum="0000000100000000000000000000000000000000" diskchecksum="0000000000000000000000000000000000000000" error_label="none"

Basically I'd like the string that starts with error: unable and ending with ] character to be removed completely so instead of :

diskchecksum="bb2a2ea700000000000000000000000000000000" error_label="nerror: unable to retrieve file meta data from cluster.local:1095 [ status=(null) ]
one"

I will have:

diskchecksum="bb2a2ea700000000000000000000000000000000" error_label="none"

I have tried the following :

sed -e 's/error:.*]$//g'

However that gives me:

diskchecksum="bb2a2ea700000000000000000000000000000000" error_label="n
one"

How do I get it to remove the newline as well when its removing the bad string?

Thanks

CodePudding user response：

With your shown samples, please try following awk using awk's RS as null here. Written and tested in GNU awk here.

awk -v RS="" '{gsub(/error: unable[^]] ]\n*/,"")} 1' Input_file

Explanation: Simple explanation would be, using global substitution to substitute error: unable till ] till newlines(0 or more occurrences) with NULL and perform printing then.

CodePudding user response：

Using gnu sed you can do this:

sed '/error: unable.*/ {s///;N;s/\n//;}' file

Or using awk:

awk 'sub(/error: unable.*/, "") {s = $0; getline; print s $0}' file

CodePudding user response：

Using sed

$ sed '/nerror:/{s/\(error_label=\)"nerror: unable[^]]*]/\1"none"/g;n;d}' input_file

CodePudding user response：

With GNU sed for -E (to enable EREs) and -z (to read the whole file at once and so allow us to match newlines in the regexp):

$ sed -Ez 's/error: unable[^]] ](\r?\n)?//g' file
path="/a/b/c" fxid="05200f4d" size="12550144" nrep="2" checksumtype="adler" checksum="045a6aa400000000000000000000000000000000"
nrep="01" fsid="132" host="cluster.local:1095" fstpath="/data/00019507/3dcd7e00" size="4574568" statsize="45745nrep="00" fsid="37" host="cluster.local:1095" fstpath="/data/000021ca/0527e888" size="12550144" statsize="12550144" checksum="bb2a2ea700000000000000000000000000000000" diskchecksum="bb2a2ea700000000000000000000000000000000" error_label="none"
path="/a/b/b98d6d3a-5c77-4223-9601-9294c73e00f9.bin" fxid="05200f4d" size="12550144" nrep="2" checksumtype="adler" checksum="045a6aa400000000000000000000000000000000"
nrep="01" fsid="36" host="cluster.local:1095" fstpath="/data/00002196/05200f4d" size="12550144" statsize="12550144" checksum="045a6aa400000000000000000000000000000000" diskchecksum="045a6aa400000000000000000000000000000000" error_label="none"
path="/a/b/c/.mb6589013703229118680.txt" fxid="0524071a" size="0" nrep="2" checksumtype="adler" checksum="0000000100000000000000000000000000000000"
nrep="00" fsid="196" host="cluster.local:1095" fstpath="/data/000021b0/0524071a" size="0" statsize="0" checksum="0000000100000000000000000000000000000000" diskchecksum="0000000000000000000000000000000000000000" error_label="none"

The above accommodates you having newlines at the end of the text you want to match whether they're just \ns or \r\ns.