Shell Programming and Scripting

hi all
i want help in sortng date in paragraphs within file ,
i want to ask as if there any option to sort a certain pattern of file not the rest of file.i.e the data of file become sorted with respect to date
i have a log file as follows

!! *A0628/081 /08-01-10/13 H 52/N=5524/TYP=INC/CAT=ID/EVENT=MAL
/NCEN=MULCT /AFUR =URAD- 13/AGEO=BAGRIAN -B01
/TEXAL=FIRE DETECT FAIL
!! *A0628/320 /08-01-15/14 H 33/N=7444/TYP=COM/CAT=ID/EVENT=MAL
/NCEN=MULCT /AFUR =URAD- 2/AGEO=CONTAIN2-B01
/TEXAL=FAULTY UNIT/COMPL.INF:
* #F0612/T11F14/NCEN=MULCT /08-01-30/11 H 41/NAM=ODCAB /TDA=0001
/N=7444/NIV=2/ENS=006/SENS=016-000-000/P='0018'H/CN=05
/EM: AFUR =URAD- 2


!! *A0628/538 /07-12-17/15 H 52/N=7576/TYP=ICT/CAT=ID/EVENT=MAL
/NCEN=MULCT /AM =SMTA1/AGEO=S1-TR01-B03-A085-R000
/TEXAL=LCL MFM SYN/COMPL.INF:
/AF=URMA1
/ ICTRQ AGCA=S1-TR01-B03-A085-R133
/AMET=01-26-03
/AFLR=217-06/CRC=NACT
!!! *A0628/294 /07-12-17/15 H 46/N=7512/TYP=SRE/CAT=IM/EVENT=MAL
/NCEN=MULCT /OBJCT=PS/AGEO=CDC
/TEXAL=SP INACCESSIBLE/COMPL.INF:
/PS =00009999/TYR=RN


!! *A0628/361 /07-12-17/15 H 46/N=7513/TYP=COM/CAT=ID/EVENT=MAL
/NCEN=MULCT /AM =SMMA/AGEO=S1-TR02-B05-A109-R000
/TEXAL=SAI-HARDWARE/COMPL.INF:
/N=3485/ALARMA MATERIAL /R=00000/CU1A
/CB 200 BLOS
! *A0628/359 /07-12-17/15 H 46/N=7514/TYP=COM/CAT=SI/EVENT=MAL
/NCEN=MULCT /AM =SMMA/AGEO=S1-TR02-B05-A109-R000
/TEXAL=SAI-SOFTWARE/COMPL.INF:
/N=4543/ABNT X25 INDISPO. /R=00083/CU1A
/SNPA=A13381/PHYSLINE=LIGP13F1


!!! *A0628/080 /08-01-13/07 H 30/N=6540/TYP=INC/CAT=IM/EVENT=MAL
/NCEN=MULCT /AFUR =URAD- 10/AGEO=RANGPUR -B01
/TEXAL=FIRE
!! *A0628/081 /07-12-29/16 H 22/N=5052/TYP=INC/CAT=ID/EVENT=MAL
/NCEN=MULCT /AFUR =URAD- 12/AGEO=KHOTYWAL-B01
/TEXAL=FIRE DETECT FAIL
!! *A0628/081 /08-01-10/13 H 52/N=5526/TYP=INC/CAT=ID/EVENT=MAL
/NCEN=MULCT /AFUR =URAD- 9/AGEO=BASTGLZR-B01
/TEXAL=FIRE DETECT FAIL


!!! *A0628/080 /08-01-10/13 H 52/N=5527/TYP=INC/CAT=IM/EVENT=MAL
/NCEN=MULCT /AFUR =URAD- 7/AGEO=ADABOSAN-B01
/TEXAL=FIRE
!!! *A0628/087 /08-01-16/01 H 28/N=7648/TYP=CLI/CAT=IM/EVENT=MAL
/NCEN=MULCT /AFUR =URAD- 10/AGEO=RANGPUR -B01

i want to extract the date , time , NCEN , EVENT , TAXAL & AGEO
i used the code following

Code:

#!/bin/bash
sed  '/^!/i\
' log | sed -n -e '/^!!! /,/^$/w critical.log' -e '/^!! /,/^$/w major.log' -e'/^! /,/^$/w minor.log'
awk 'BEGIN {FS="/"; RS=""} { printf "%s/%s/%s/\n%s\n%s\n%s\n%s/%s\n\n", $1, $3, $4, $10, $12, $8, $14, $15}' minor.log 
exit 0

this code does seperate the paragraphs follwing ! , !! , !!! (i.e. minor , major & critical alarms of the log file respectively ,in their respective log files i.e. minor.log ,major.log & critical.log )
and 'awk' gives me output as following

! *A0628/07-12-17/15 H 58/
NCEN=MULCT
AGEO=S1-TR01-B03-A085-R000
EVENT=MAL
TEXAL=AIS/COMPL.INF:/ /AF=URMA1

but i want the output to be sorted by date
(date pattern=/07-12-17/)
i tried to use the sort -k but i coldnt understand the right pattern to use it
or I have to sort this date pattern by loops ?
cant understand???
i tried to sort the date pattern by reg exp and to some extent i suceeded by i ca only compare two dated i couldnt get the logic to make the loop to check each date in file and then puxh the latest date entry on top each time
i used the code

Code:

#!/bin/bash
read date1
read date2
a=${date1:0:2}
b=${date2:0:2}
if $a -ge $b
then
echo $a
elif $a -le $b
then echo $b
else
echo "sorry"
fi
exit 0

plz do it if u can or tell me the right way to go ahead

extract the fields first then sort them:

Code:

# extract date , time , NCEN , EVENT , TEXAL & AGEO, sort by date
# remove trailing blanks and then remove blank lines, then read, format line, print.
sed 's/[\t ]$//g' filename | grep -v '^$' | \
awk -F'/' ' {if(index($1, "!") == 1)
                { printf ("\n%s %s ", $3, $4)}
            else
                {   for(i=1; i<=NF; i++)
                	{
                	    if(index($i,"NCEN") == 1 || index($i,"EVENT")== 1 ||
                	       index($i,"TEXAL") == 1 || index($i,"AGEO") == 1 )
                	       { printf("%s ", $i) }
                	}
                }
             } END { printf("\n") } '  | sort

Is each record separated by a newline?? I can't tell, because sometimes the records have two newlines.

actually the line pattern is not regular but i extracted the output data
from file i.e. date , time ,NCEN , TAXAL , EVENT ,& AGEO
now i have the regular 1 line space in output

! *A0628/07-12-17/15 H 58/
NCEN=MULCT
AGEO=S1-TR01-B03-A085-R000
EVENT=MAL
TEXAL=AIS/COMPL.INF:/ /AF=URMA1

i used the following code to extract

Code:

#!/bin/bash
sed  '/^!/i\
' log | sed -n -e '/^!!! /,/^$/w critical.log' -e '/^!! /,/^$/w major.log' -e'/^! /,/^$/w minor.log'
awk 'BEGIN {FS="/"; RS=""} { printf "%s/%s/%s/\n%s\n%s\n%s\n%s/%s\n\n", $1, $3, $4, $10, $12, $8, $14, $15}' minor.log 
exit 0

now i need to sort it by date
but couldnt find the way

a reply. see if you can use it.

Dude, that's way overkill. I guess for python programmers, every problem is a nail?

nabmufti,

The normal sort routine will work. It looks like the date field is the third one within slashes, correct? If so, we can just do this:

sort -t / -k 3,4

Pipe the output of your script into this sort, and that should be it! Now, this works because the date is in proper lexical order (YY-MM-DD) and doesn't require individual sorting of the number space. If you had DD-MM-YY, it would be a tad bit trickier.

If you have multiple events on the same day, an you want to sort on another field, you can add a -k parameter:

sort -t / -k 3,4 -k 2,3

That will sort on the second /-separated field.

depends on how you look at it.
using that many a combination of sed's, grep's , awk's and sort, IMO, is the overkill.