Sed command for handling string replacements by matching the pattern

I am create a command using sed, which should work for these three scenarios.

echo "a_0001_hst_script5v00.sh 06/09/2023 06:30:30 06/09/2023 06:32:45 SU 307404729/1 0" |sed -E "s#([^/]*)?([0-9]{2}/[0-9]{2}/[0-9]{4})#echo $(date -d '\' +'%d-%^h-%Y')#g;s/echo //2ge"

is producing the desired result as below.

a_0001_hst_script5v00.sh 09-JUN-2023 06:30:30 09-JUN-2023 06:32:45 SU 307404729/1 0

essentially its reprinting the given input record while converting the date to DD-MMM-YYYY format as shown in the result above.

but there is slight variation of the input record where second date is blank (-----). so to process that record format, i had to modify the command slightly as shown below.

echo "a_0001_hst_script5v00.sh 06/09/2023 06:30:30 ----- RU 307404729/1 0" |sed -E "s#([^/]*)?([0-9]{2}/[0-9]{2}/[0-9]{4})#echo $(date -d '\' +'%d-%^h-%Y')#g;e"

and produced the desired result as below one.

a_0001_hst_script5v00.sh 09-JUN-2023 06:30:30 ----- RU 307404729/1 0

and there is a third scenario where the record format having the both dates go blank like shown below.

a_0001_hst_script5v00.sh ----- ----- OH 307404729/1 0

and it should just reprint this 3rf format record as it is.

i have been trying to write this command in single line which works for al these 3 scenarios of the record formats given above and struggling.

could some one throw some light on this?

You can change where the command is executed i.e the first pass before removing the duplicate echo and change the condition for the removal of additional echo

$ echo "a_0001_hst_script5v00.sh 06/09/2023 06:30:30 06/09/2023 06:32:45 SU 307404729/1 0
a_0001_hst_script5v00.sh 06/09/2023 06:30:30 ----- RU 307404729/1 0
a_0001_hst_script5v00.sh ----- ----- OH 307404729/1 0" | \
sed -E "s#([^/]*)?([0-9]{2}/[0-9]{2}/[0-9]{4})#echo $(date -d '\' +'%d-%^h-%Y') #ge;s/echo //g"
a_0001_hst_script5v00.sh 09-JUN-2023 06:30:30 09-JUN-2023 06:32:45 SU 307404729/1 0
a_0001_hst_script5v00.sh 09-JUN-2023 06:30:30 ----- RU 307404729/1 0
a_0001_hst_script5v00.sh ----- ----- OH 307404729/1 0

With your shown samples please try following awk code. Written and tested in GNU awk should work in any version.

awk '
BEGIN{
  split("Jan,Feb,Mar,Apr,May,Jun,Jul,Aug,Sept,Oct,Nov,Dec",months,",")
}
{
  line=$0
  while(match(line,/[0-9]{2}\/[0-9]{2}\/[0-9]{4}/)){
    val=substr(line,RSTART,RLENGTH)
    split(val,arr,"/")
    val2=arr[2]"-"months[sprintf("%01d",arr[1])]"-"arr[3]
    sub(val,val2)
    line=substr(line,RSTART+RLENGTH)
  }
}
1
'  Input_file

This might work for you (GNU sed):

sed 's/ ..\/..\/.... / $(date -d & +"%d-%^h-%Y") /g;s/.*/echo "&"/e' file

Replace each date with a command substitution and finally evaluated the echoed line.

Alternative using only sed:

sed -E 's/$/\n01JAN02FEB03MAR04APR05MAY06JUN07JUL08AUG09SEP10OCT11NOV12DEC/
        :a;s/ (..)\/(..)\/(....) (.*\n.*(...))/ -- /;ta;P;d' file