search with grep using 3 criteria

[future_chris]

Hi guys,

I have a question about using grep to search a document. Now the grep command I am trying to use is:

# fsmedinfo 000023 -l | grep -e (1) -e dr_policy -e ((001)) >> text

What I am trying to do is scan a builtin command with a qantam tape backup drive. This command will display a tape number for example 000023(1) or 000024(2) and then a tape policy either dr_policy or archive_policy and finally if media on the tape has been unlinked then it will display a ((001)) at the start of a file list.

So what I want to do with my script is to find all the tapes in a predefined range and search for 1st version tapes (1), all tapes in the dr_policy and finally any tape that has unlinked media ((001)). If this is not true for all of my criteria then I don't want grep to output to my text file.

I have been able to get the above command to work but if it finds a 2nd version tape (2) but there is unlinked media ((001)) and has a Policy of dr_policy then i get that information written to my text file which i don't want.

Is there a way with grep or should I use another command? I was looking into using awk but have not got very far with it.

Example Output I want to filter:

-------------------------------------------------------------------------------
Media Information Report Wed Mar 31 11:52:33 2010
Media ID: 000025(1)
Media Type: LTO
-------------------------------------------------------------------------------
Storage Area: VolSub
Class ID: archive_pc Bytes Used: 869,902,569,216
Last Accessed: 20-mar-2010 02:17:29 Space Remaining: 0
Media Status: AVAIL Percent Used: 100.00
Write Protect: N Suspect Count: 0
Mark Status: UNMARKED Mount Count: 18
Medium Location: SLOT/BIN
Formatted: Y
Number of Segments: 5809
External Location: N/A

[meton_magis]

I'm not quite understanding what you're looking at, however, if you are trying to use grep to filter content on 2 different lines, it will never work. that is the number 1 limitation of grep, is that it does not span different lines. Awk is good, however I found that it is better to focus on learning perl. Perl is installed everywhere on unix (almost) and can do everything awk does. And there is 1 official version of perl. Awk you have old awk, new awk, gnu awk, solaris awk, AIX awk, HP-UX Awk ... too confusing.

I just use perl now, it works pretty well.

[future_chris]

Hi Meton_magis

Thanks for the reply, what I am ultimately wanting to do is use grep if possible to search the pipped output of my fsmedinfo command and if it finds (1) and dr_policy and ((001)) then I want it to output via output redirection to a text file the following:

Media ID: 000025(1)
Class ID: dr_policy
((001)) 643657547 /user/media/movie.mov
((001)) 567746575 /user/media/movie2.mov

If I was to just grep -i '(1)' using my storage media command fsmedinfo then i would get an output of:

#Media ID: 000025(1) (if there was a (1) at the end of the Media ID field.)
#

As there is no other instance of (1) except in the Media ID field I get my desired output.

If then i use grep -e '(1)' -e dr_policy and -e '((001))' and all of the following patterns are true then I get the expected output of grep (as above). If for any reason one of the 3 criteria is not true then I still get a result but grep will only output the lines that match what is true out of the 3 criteria i have put in the command.

This result i don't want. Can't i tell grep that if all 3 patterns are not found in the output of my command to just stop and return an exit status of 1?

Example:
#grep -e <CRITERIA 1> AND -e <CRITERIA 2> AND -e CRITERIA 3>

I hope this helps clarify my scripting dilemma. But I am investigating Perl as well.