grep a file process the output to variable-CodePudding

I have a file with under 200 entries, they are not really uniform but with a grep before and after command I can get client data. Here is what the output of the command
grep -i -b -A4 " Type : Client" sample.txt

Index     : 11
Type      : Client, Disk Proxy Client
Location  : 192.1680.11
Up Time   : 0d 0h 25m
Revision  : 6.3.1

Index     : 12
Type      : Client
Location  : 192.1680.12
Up Time   : 0d 0h 25m
Revision  : 7.0.1

Index     : 56
Type      : Client
Location  : 10.0.10.112
Up Time   : 0d 0h 25m
Revision  : 5.3.1

I want to run a while/for loop and separate each line into a variable to then push these values to database.

As a sample of getting I can get the "Location" for each grep result I can use this. I would like to put this in a script and grab all the output in to individual variables.
cat sample.txt |grep -i -b -A4 "Type" |while read client; do echo "$client" |grep Location |awk '{print $4}'; done

192.168.0.11
192.168.0.12
10.0.10.112

But if I try to put the same command into a multi-line script it gets processed as individual lines.
I would like to do more grep/sed..etc to get

index=11
type=client
location=192.168.0.11

fake sql query
sql -update .... $index,$type,$location

CodePudding user response：

It seems like this might be what you're trying to do:

$ cat tst.sh
#!/usr/bin/env bash

while IFS= read -r rec; do
     echo sql -update .... "$rec"
done < <(
    awk '
        BEGIN {
            FS = "[[:space:]]*:[[:space:]]*"
            OFS = ","
        }
        $1 ~ /^(Index|Type|Location)/ {
            rec = rec sep $2
            sep = OFS
        }
        !NF {
            print rec
            rec = sep = ""
        }
        END {
            print rec
        }
    ' "${@:--}"
)

$ ./tst.sh sample.txt
sql -update .... 11,Client, Disk Proxy Client,192.1680.11
sql -update .... 12,Client,192.1680.12
sql -update .... 56,Client,10.0.10.112

or if you really want 3 separate shell variables for some reason then:

$ cat tst.sh
#!/usr/bin/env bash

while IFS=':' read -r index type location; do
     echo sql -update .... "$index,$type,$location"
done < <(
    awk '
        BEGIN {
            FS = "[[:space:]]*:[[:space:]]*"
            OFS = ":"
        }
        $1 ~ /^(Index|Type|Location)/ {
            rec = rec sep $2
            sep = OFS
        }
        !NF {
            print rec
            rec = sep = ""
        }
        END {
            print rec
        }
    ' "${@:--}"
)

Remove the echo and fix the sql query if I'm right, otherwise update your question with more complete requirements and more truly representative sample input/output.

Note that the above is directly using sample.txt as input, not the output of your existing grep command, we're no longer using that.

CodePudding user response：

If sed is an option and your expected output is correct (lower case as well as client being chopped at the ,), you can try this;

$ sed '/Up/{N;d};s/\([[:alnum:]]\ \)[^[:alnum:]]*\([^,]*\).*/\L\1=\2/' file
index=11
type=client
location=192.1680.11

index=12
type=client
location=192.1680.12

index=56
type=client
location=10.0.10.112