bash sed如何追加到行首，与行首匹配的变量

Question

I'm having some trouble figuring out how to append at the beginning of each line if the beginning of the line matches the servername in a variable. Although nice to have but not required is if no site was matched up, to indicate "unknown-site" at the beginning of line.

I do have a known list of machine names for each site.

site1="server1 server2 ..."
site2="server3 ..."

I was thinking the code should be something like

for each servername in $site1 do; sed command in file
for each servername in $site2 do; sed command in file

The raw data that was supplied to me

server1 data1...
server1 data2...
server3 data3...
server3 data2...
server2 data4...
server3 data3...
server2 data5...

And this should be the out put

site1 server1 data1...
site1 server1 data2...
site2 server3 data3...
site2 server3 data2...
site1 server2 data4...
site2 server3 data3...
site1 server2 data5...

Answer 1

If you can arrange the server<>site mapping to one file ( servers ) like this:

server1 site1
server2 site1
server3 site2

Then this becomes easy to do with awk :

$ awk 'FNR==NR {s[$1] = $2; next} {print s[$1], $0}' servers in
site1 server1 data1...
site1 server1 data2...
site2 server3 data3...
...

NR is the number of records (lines) seen by awk this far, FNR the number of records within this file seen this far. Comparing them with FNR==NR is a common idiom for separating the first file from the rest. So, for each line of the first file, we set the values found to the array s , with the first field ( $1 ) as key, and the second ( $2 ) as value, then jump to next line. (All arrays in awk are associative, and work with strings as keys/indexes.) Now we have the site names matching the servers in array s (eg s["server1"] == "site1" ) and for the subsequent files, we just use the first field as an index to the array, print the value found there, along with the full input line ( $0 ).

We could also add a known marker (instead of empty) for the unknown ones by checking if s[$1] exists and is nonempty:

$ awk 'FNR==NR {s[$1] = $2; next} {if (s[$1]) print s[$1], $0; else print "unknown", $0;  next}' servers data.txt

With the server names as you had in the shell, you could produce the servers file with something like this:

site1="server1 server2 ..."
site2="server3 ..."
true > servers        # to clear the file
for x in $site1 ; do echo "$x site1" ; done >> servers
for x in $site2 ; do echo "$x site2" ; done >> servers

Answer 2

Try this:

awk -F '[ ="]+' 'NR==FNR {for (i=2; i<=NF; i++) { sites[$i]=$1; }; next }
                 { print sites[$1] " " $0 }' sites data