I am trying to pull a value from a sequence, based on the entry of the parameter in the file. In my case, the values are going to be either "seq", "date", or "count".
function ParseHeader {
local varHeaderLine=$1
local varType=$2
local varOut=""
if [[ -z "$varType" ]]; then
echo "Cannot find! - ($varType)"
exit 1
elif [[ "$varType" = "seq" ]]; then
varOut=$(echo "$varHeaderLine" | cut -d\| -f 3)
echo "Sequence: $varOut"
elif [[ "$varType" = "date" ]]; then
varOut=$(echo "$varHeaderLine" | cut -d\| -f 4)
echo "Date: $varOut"
elif [[ "$varType" = "count" ]]; then
varOut=$(echo "$varHeaderLine" | cut -d\| -f 5)
echo "Line Count: $varOut"
else
echo "Invalid input! Your Header is: $varHeaderLine"
fi
return $?
}
sequence='XXXXX|Y|1|20220429|9999'
ProfSeq=$(ParseHeader "${sequence}" "count")
echo $ProfSeq
This is a piece out of the whole file, but what is happening here is that I am going to receive a file, pull its sequence, and then pull some value from that particular sequence. Thanks for the help, this is what I ended up with. Now I'm looking for it to be less CPU intensive, if possible.
Note: This piece is out of a while file script that will be running daily for different servers in different environments
CodePudding user response:
Focusing solely on the current piece of code ...
Assumptions/Understandings:
- the variable
ProfSeq
is to contain the output from the function'secho
calls; we can eliminate the subprocess call (ProfSeq=$( ... subprocess call ...)
) by having the function replace theecho "..."
calls withProfSeq="..."
(ie, have the function directly populate theProfSeq
variable - additional performance improvements can be derived from eliminating the double subprocess calls to split
varHeaderLine
($(echo ... | cut ...)
) with thebash/read
builtin
One idea for a function rewrite:
ParseCOTHeader () {
local varHeaderLine=$1
local -l varType=$2 # force to all lowercase
local ignore1 ignore2 in_seq in_date in_count rc
# split varHeaderLine into 5 variables based on '|' delimiter:
# NOTE: assumes varHeaderLine has exactly 5x '|' delimited fields otherwise OP
# may want to add some code to validate the structure of varHeaderLine
IFS='|' read -r ignore1 ignore2 in_seq in_date in_count <<< "${varHeaderLine}"
# reset variables:
ProfSeq=
rc=0
if [[ -z "${varType}" ]]
then
ProfSeq="Cannot find! - ($2)" # since varType=$2 is empty at this point we will always print '()' ... ?
rc=1
else
case "${varType}" in
seq) ProfSeq="Sequence: ${in_seq}" ;;
date) ProfSeq="Date: ${in_date}" ;;
count) ProfSeq="Line Count: ${in_count}" ;;
*) ProfSeq="Invalid input! Your Header is: '$varHeaderLine'"
rc=1 ;;
esac
fi
return "${rc}"
}
Putting the function through its paces:
$ sequence='SWHHR|H|1|20220429|9999'
$ ParseCOTHeader "${sequence}" 'seq'
$ echo "$? : ${ProfSeq}"
0 : Sequence: 1
$ ParseCOTHeader "${sequence}" 'DaTe'
$ echo "$? : ${ProfSeq}"
0 : Date: 20220429
$ ParseCOTHeader "${sequence}" 'COUNT'
$ echo "$? : ${ProfSeq}"
0 : Line Count: 9999
$ ParseCOTHeader "${sequence}" 'fire-truck'
$ echo "$? : ${ProfSeq}"
1 : Invalid input! Your Header is: 'SWHHR|H|1|20220429|9999'
$ ParseCOTHeader "${sequence}" ''
$ echo "$? : ${ProfSeq}"
1 : Cannot find! - ()
If repeated calls to the function need to save the message string in different variables:
- add the line
local -n myoutput=$3
in the top section of the function (this defines a local nameref variablemyoutput
to act as a pointer to the variable name passed in $3) - replace all
ProfSeq="..."
withmyoutput="..."
in the function - remove the line
ProfSeq=
(variable reset) in the function
Taking the new function for a test spin:
$ sequence='SWHHR|H|1|20220429|9999'
$ ParseCOTHeader "${sequence}" 'seq' ProfSeq
$ echo "$? : ${ProfSeq}"
0 : Sequence: 1
$ ParseCOTHeader "${sequence}" 'DaTe' ProfSeq
$ echo "$? : ${ProfSeq}"
0 : Date: 20220429
$ ParseCOTHeader "${sequence}" 'COUNT' ProfSeq
$ echo "$? : ${ProfSeq}"
0 : Line Count: 9999
$ ParseCOTHeader "${sequence}" 'fire-truck' SomeOtherVar
$ echo "$? : ${SomeOtherVar}"
1 : Invalid input! Your Header is: 'SWHHR|H|1|20220429|9999'
$ ParseCOTHeader "${sequence}" '' YetAnotherVar
$ echo "$? : ${YetAnotherVar}"
1 : Cannot find! - ()
CodePudding user response:
You could use read
and associative arrays:
$ function ParseHeader {
local -A fields=(["seq"]=2 ["date"]=3 ["count"]=4)
local -A mess=(["seq"]="Sequence" ["date"]="Date" ["count"]="Line Count")
IFS='|' read -ra arr <<< "$1"
if [[ -n "${fields[$2]: x}" ]]; then
printf -v ProfSeq '%s: %s' "${mess[$2]}" "${arr[${fields[$2]}]}"
else
printf -v ProfSeq 'Invalid input! Your Header is: %s' "$1"
fi
}
$ ParseHeader "XXXXX|Y|1|20220429|9999" "seq"
$ echo "$ProfSeq"
Sequence: 1
$ ParseHeader "XXXXX|Y|1|20220429|9999" "date"
$ echo "$ProfSeq"
Sequence: 20220429
$ ParseHeader "XXXXX|Y|1|20220429|9999" "count"
$ echo "$ProfSeq"
Line Count: 9999
$ ParseHeader "XXXXX|Y|1|20220429|9999" "foo"
$ echo "$ProfSeq"
Invalid input! Your Header is: 31