I have the following output returned by awk:
qh1adm 20130711151154 : tp import all QH1 u6 -Dsourcesystems=BFI,EBJ
qh1adm 20130711151155 : tp import all QH1 u6 -Dsourcesystems=E7B,B17
qh1adm 20130711151200 : tp import all QH1 u6 -Dsourcesystems=BFI,EBJ
I want to store this in an array in-order to count the number of records. If it is more than 1 only then I want to display the contents of this array. Can you please assist?
Thanks a lot.
Using bash/zsh, capture the output, count the lines, and print it if it meets your criteria:
output=$(awk 'something that returns what you showed in your question')
nlines=$(wc -l <<< "$output")
if ((nlines > 1)); then
echo "$output"
fi
awk '{s[c++]=$0} END { if (c>1) { for (c in s) { print s[c]; } } }'
yourawkcommand | awk ....... | awk 'NR == 1 {first=$0} NR == 2 {print first} NR > 1 {print}'
This will output nothing if there are 1 or fewer lines in the input. Otherwise it prints the entire input. It's efficient because it stores only the first line, so it does not waste memory on storing the entire input in an array.
You could just store the output of awk in a file and then count the lines of the file.
TEMP_FILE=`mktemp`
awk_command > $TEMP_FILE
if [[ `wc -l < $TEMP_FILE` -gt 1 ]]
then
cat $TEMP_FILE
fi
rm $TEMP_FILE
In case anyone in future is reading this and wondering how to actually get multi-line awk output into a bash array:
$ cat ./tst.sh
# ensure globbing is off and set IFS to a newline after saving original values
oSET="$-"; set -f; oIFS="$IFS"; IFS=$'\n'
array=( $(
awk 'BEGIN{
print "qh1adm 20130711151154 : tp import all QH1 u6 -Dsourcesystems=BFI,EBJ"
print "qh1adm 20130711151155 : tp import all QH1 u6 -Dsourcesystems=E7B,B17"
print "qh1adm 20130711151200 : tp import all QH1 u6 -Dsourcesystems=BFI,EBJ"
}'
) )
# restore original IFS and globbing values
IFS="$oIFS"; set +f -"$oSET"
numElts="${#array[@]}"
if (( numElts > 1 ))
then
for (( i=0; i < numElts; i++ ));
do
printf "array[%d]: %s\n" "$i" "${array[$i]}"
done
fi
$
$ ./tst.sh tmp
array[0]: qh1adm 20130711151154 : tp import all QH1 u6 -Dsourcesystems=BFI,EBJ
array[1]: qh1adm 20130711151155 : tp import all QH1 u6 -Dsourcesystems=E7B,B17
array[2]: qh1adm 20130711151200 : tp import all QH1 u6 -Dsourcesystems=BFI,EBJ
but apparently none of that is actually necessary in this case and the whole thing should probably just be done within the one awk command.
The above has been updated to address issues raised by and input from @CharlesDuffy below.
array=( $(stuff) ) does string-splitting and glob expansion on the output of stuff, making it hard to guarantee that results come back exactly the way you want them. A safer approach would be to have awk emit NUL-delimited entries, and to use array=(); while read -r -d ''; do array+=( "$REPLY" ); done < <(awk ...) to read them in.* in the output from awk; it would be replaced with a list of files in the current directory before being put into array.set -f to disable globbing, and set +f to reenable it later. That's somewhat bad form for the same reason changing IFS is (without explicitly scoping that IFS change to a single command, as done by IFS=... read) -- you're increasing the amount of context someone needs to have in their head to understand what a given piece of script does.$- before the set -f and restore it later. Obviously I can also save IFS and restore that to it's original value later too. Let me update the script now then please take another look and see if there's anything else. How about word-splitting, is that really an issue in this case?