sed hash in multiple lines

Shell Programming and Scripting
1

Hi everybody,

I am writing a little script to manage keystores and need some help with sed.

The output of the keystore file is like:

vi 2, Dec 7, 2012, trustedCertEntry,
Certificate fingerprint (MD5): F9:1F:FE:E6:A3:CB:99:88:44:D4:67:ED:G5:F8:97:7A
system@remote-server, Dec 17, 2013, PrivateKeyEntry,
Certificate fingerprint (MD5): 2A:DD:4B:C5:AD:B0:MF:96:51:B0:14:4A:B5:6C:99:FB
verisign intermediate 1, Dec 7, 2012, trustedCertEntry,
Certificate fingerprint (MD5): 4E:F6:A0:FC:40:0C:AE:9C:05:2F:AE:99:C6:7D:37:9F

So what I want, is the hash sum of "system@remote-server" (e.g. 2A:DD:4B:C5:AD:B0:MF:96:51:B0:14:4A:B5:6C:99:FB )

What I tried as far is:

sed -e "system@remote-server.*PrivateKeyEntry,/,/Certificate fingerprint.*:[a-zA-Z0-9]{2}$/ {/\([[a-zA-Z0-9]{2}[:]]*[a-zA-Z0-9]{2}$\)/\1/p}"

and

sed -n "s/system@remote-server.*PrivateKeyEntry.*Certificate fingerprint:*([a-zA-Z0-9_-]): \([[a-zA-Z0-9]{2}[:]]*[a-zA-Z0-9]{2}$\)/\1/p"

but I don't get any output for them.

ps: the easy way would be grep -A 1 "system@remote-server.*PrivateKeyEntry" | tail -1 | cut -f 4 -d " " , but it is not reliable. Therefore I want to do it with sed.

Does anybody have a solution for me?

Thank you all in advance and regards
Hamss

2

Hello,

Could you please use the following code for same.

awk '/system@remote-server, Dec 17, 2013, PrivateKeyEntry,/ {getline; print $0}' file_name | sed 's/Certificate fingerprint (MD5): //g'

Output will be as folllows.

2A:DD:4B:C5:AD:B0:MF:96:51:B0:14:4A:B5:6C:99:FB

Thanks,
R. Singh

1 Like
3

Hi,
In sed:

sed -n -e '/system@remote-server.*PrivateKeyEntry/,/Certificate fingerprint/{/system@remote-server.*PrivateKeyEntry/d;s/.* \([0-9A-F]\+:\)/\1/p;}'

Simplified:

sed -n -e '/system@remote-server.*PrivateKeyEntry/,/Certificate fingerprint/{/Certificate fingerprint/s/.* \([0-9A-F]\+:\)/\1/p;}'

Because "d" command is not need with "-n" option...(why delete line that we won't print ?)
Regards.

1 Like
4 
sed -n '/system@remote-server/{n;s/.* //p}' file

2A:DD:4B:C5:AD:B0:MF:96:51:B0:14:4A:B5:6C:99:FB
1 Like
5

I'm not sure that this solution is possible otherwise the solution will also be possible with grep...

6

This actually does not work in my case...

Thank you all for the suggestions. I need to play a bit around with them.

General question; Is it possible to merge multiple regex into one?

E.g. for this case.
We have a this hash and want to make a regex ohny for this pattern 2A:DD:4B:C5:AD:B0:MF:96:51:B0:14:4A:B5:6C:99:FB

So I am increasing the patterns, and merging one into an other:

  1. pattern: ([A-Z0-1][A-Z0-1]:) set pattern
  2. pattern: ([[A-Z0-1][A-Z0-1]:]*) pattern 1 0-infinite times
  3. pattern: ([[A-Z0-1][A-Z0-1]:]*[A-Z0-1][A-Z0-1]) and put the last two chars at the end

Is this possible?

---------- Post updated at 07:19 PM ---------- Previous update was at 06:05 PM ----------

Thanks this works good in my case. But I would like to pattern the hash itself. How would you do that?

7 
$ cat file
vi 2, Dec 7, 2012, trustedCertEntry,
Certificate fingerprint (MD5): F9:1F:FE:E6:A3:CB:99:88:44:D4:67:ED:G5:F8:97:7A
system@remote-server, Dec 17, 2013, PrivateKeyEntry,
Certificate fingerprint (MD5): 2A:DD:4B:C5:AD:B0:MF:96:51:B0:14:4A:B5:6C:99:FB
verisign intermediate 1, Dec 7, 2012, trustedCertEntry,
Certificate fingerprint (MD5): 4E:F6:A0:FC:40:0C:AE:9C:05:2F:AE:99:C6:7D:37:9F

$ awk '/system@remote-server/ && getline && gsub(/.*\): /,x)' file

2A:DD:4B:C5:AD:B0:MF:96:51:B0:14:4A:B5:6C:99:FB
1 Like
8

Some more:

awk 'f{print $NF; f=0} /system@remote-server/{f=1}' file

sed '/system@remote-server/!d;N;s/.* //' file
2 Likes
9

Hello Scrutinizer,

Thanks a lot for nice code. Could you please explain the code please.

Thanks,
R. Singh

10

Sure:

awk '
  f{                         # if f=1 then (if a match was found previously)
    print $NF                # print the last field
    f=0                      # reset f ( set found is false)
  } 
  /system@remote-server/{    # if a match is found
    f=1                      # then set f to 1
  }
' file

sed '
  /system@remote-server/!d   # if a match is found then do not delete the line
  N                          # append the next line to the current line
  s/.* //                    # delete everything unto and including the last space
' file
1 Like
11

In your hash, I doubt you have :MF:
In general, it's a serie of hex digit.
Your pattern 2 is not pattern 1 0-infinite times, you can simplified your pattern 2 as: ([A-Z0-1:]*)
If you want pattern 1 0-infinite, you can do as: (([A-Z0-1][A-Z0-1]:)*)
And so, pattern 3: (([A-Z0-1][A-Z0-1]:)*[A-Z0-1][A-Z0-1])

And if you want that your pattern is a regex for your hash, the pattern 3 will:
(([0-9A-F][0-9A-F]:)*[0-9A-F][0-9A-F])

Regards.

1 Like
12

Thank you Scrutinizer for explaination. Here is one more approach for same.

$ awk '/system@remote-server, Dec 17, 2013, PrivateKeyEntry,/ {getline; gsub(/Certificate fingerprint/, "");print$NF}' file_name
2A:DD:4B:C5:AD:B0:MF:96:51:B0:14:4A:B5:6C:99:FB

Thanks,
R. Singh

1 Like
13

If you want the regex to match the fingerprint, try:

sed -rn 's/.*(([A-F0-9]{2}:){15}[A-F0-9]{2})/\1/p' file

This does not necessarily work with all sed versions, and, as pointed out before, it matches HEX numbers only - if you really have sth. like MF in there, adapt the ranges.

1 Like
14

Thanks a lot, that was what I wanted to know :b:

---------- Post updated at 01:44 PM ---------- Previous update was at 01:39 PM ----------

Thats actually true. I've changed some digs before posting. :slight_smile: