Hi,
Im using the below command to search a file for multiple expressions if the 4th expression doesnt exist the command simply lists what it can find and ignores what it cant. Is there any way to get the command to output an error or a message if it cant find the 4th expression to a file?
Any help will be much appreciated 
egrep -is 'Schema Area|stock|route_drop|test123' /apps/Main_structure.txt 1>/apps/grep_error.txt
SunOS sants 5.10 Generic_118822-30 sun4u sparc SUNW,Sun-Fire-V490
You can use an egrep -v on a pipe to remove lines that match and an egrep to save lines that match.
Here is an approach using nawk
nawk ' BEGIN {
IGNORECASE=1;
} /Schema Area/||/stock/||/route_drop/||/test123/ {
if($0~/test123/)
c++;
print;
} END {
if(c==0)
print "Error: 4th expression not found!";
}' /apps/Main_structure.txt
You can also use sed, where you can act on lines that have a pattern '/pattern/{...}' or not '/pattern/!{...}' or test one inside the other.
if grep -q "Schema Area" file &&
grep -q stock file &&
grep -q route_drop file &&
grep -q test123 file
then
echo correct
else
echo incorrect
fi
What does the input file look like?
You can do this in one pass in awk, setting flags for each pattern, discarding input and testing the flags at eof.
sorry I dont have the use of nawk or grep with the -q switch
egrep -v wouldn't get what Im looking for as it wouldnt highlight any errors
sorry I only have limited knowledge of awk and sed so Im not sure how I would use them in this case
What I am looking for is by using grep or egrep or any other command to search though a file for expression1|expression2|expression3 etc... and if it can't find one of the expressions output an error along the lines of "can't find expression1". If it does find the expression in the file then do nothing and output nothing.
Thankyou for all your help so far people
---------- Post updated at 04:05 AM ---------- Previous update was at 04:01 AM ----------
the input file looks like this and carries on:
Area Name: Schema Area
Size: 77823936, Records/Block: 64, Area Number: 6
Area Name: stock
Size: 1024192, Records/Block: 64, Area Number: 100
Area Name: route_drop
Size: 4096768, Records/Block: 256, Area Number: 101
You probably do have grep -q somewhere on your system, on Solaris use /usr/xpg4/bin/grep
Otherwise you could use:
if grep "Schema Area" >/dev/null file &&
grep stock >/dev/null file &&
grep route_drop >/dev/null file &&
grep test123 >/dev/null file
then
echo correct
else
echo incorrect
fi
or
if { grep "Schema Area" file &&
grep stock file &&
grep route_drop file &&
grep test123 file
} > /dev/null
then
echo correct
else
echo incorrect
fi
fantastic thank you Scrutinizer that option works perfectly
not sure why it wont let me use the grep -q option youve posted I get this when I try
grep: illegal option -- q
Usage: grep -hblcnsviw pattern file
/usr/xpg4/bin/grep
does exist though
Good, yes, then you could use /usr/xpg4/bin/grep -q instead of grep -q . That standard grep on Solaris is only present for legacy reasons..
That works now too both produce the same results so is one prefered to the other or doesnt it matter?
It's up to you
awk '/^size/ && ! NF==6 { print "Something missing" } filename
---------- Post updated at 05:01 PM ---------- Previous update was at 04:56 PM ----------
please ignor my previous post.
awk '$1 ~ /^Size/ && NF != 7 { print "Something missing"}' filename
Things like -q are command specific, but if you want a silent test, just run the command into a variable:
cmd_oput=$( some_command args ... 2>&1 )
---------- Post updated at 10:51 AM ---------- Previous update was at 10:38 AM ----------
You can do file evaluation in one pass in shell, too, using file globbing metachar:
ct1=0 ct2=0 ct3=0 ct4=0
while read l
do
case "$l" in
(*"Schema Area"*)
(( ct1++ ))
;;
esac
case "$l" in
(*stock*)
(( ct2++ ))
;;
esac
case "$l" in
(*route_drop*)
(( ct3++ ))
;;
esac
case "$l" in
(*test123*)
(( ct4++ ))
;;
esac
done < in_file
# process counts $ct1-4 to determine response
This one will output the missing pattern
$ echo 'Schema Area|stock|route_drop|test123' | tr '|' '\n' > tmp; grep -Eiof tmp file | grep -v -f- tmp; rm tmp
test123
And, this is DGPickett's proposal, slightly simplified:
while read l
do case "$l" in
(*"Schema Area"*) (( ct1++ ));;
(*stock*) (( ct2++ ));;
(*route_drop*) (( ct3++ ));;
(*test123*) (( ct4++ ));;
esac
done < file
Just remember that if a line has pattern 1, you will miss all other patterns on that line.
Oh ... yes, you are right. In bash, appending a & the ;; making it read ;;& would help. This may not be available in other shells.