I want to extract last part of directories from specific date=2022-01-10 and having construction like :

/rec/flux_entrant/le5/archive/tble91_formation/20220104-020124
/rec/flux_entrant/le5/archive/tble91_formation/20220105-020113
/rec/flux_entrant/le5/archive/tble91_formation/202201013-020124
/rec/flux_entrant/le5/archive/tble91_formation/202201014-020123

 ...

As you can see we have many date here 20220104 , 20220105 , 202201013 and 202201014

In other world I should develop a loop operator that take all path which exists in :

/rec/flux_entrant/le5/archive/tble91_formation/

and return a list having values like : 202201013-020124 and 202201014-020123 because 202201013 and 202201014 are dates given after 2022-01-10

Any help please ? Thank you

CodePudding user response：

This might be what you're looking for

#!/bin/bash

date='2022-01-10'

for path in /rec/flux_entrant/le5/archive/tble91_formation/*; do
    lastpart=${path##*/}
    ymd=${lastpart%%-*}
    [[ $ymd = ??????0?? ]] && ymd=${ymd:0:6}${ymd:7}
    if [[ $ymd = [0-9][0-9][0-9][0-9][0-9][0-9][0-9][0-9] ]]; then
        [[ $ymd > ${date//-} ]] && echo "$lastpart"
    else
        echo "Unrecognized date format in file $lastpart" >&2
    fi
done

CodePudding user response：

Assuming the month and day fields always contain preceding 0, would you please try:

#!/bin/bash

refdate='2022-01-10'

for f in /rec/flux_entrant/le5/archive/tble91_formation/*; do
    base=${f##*/}
    if [[ $base =~ ^([0-9]{4})(0[1-9]|01[0-2])(0[1-9]|0[1-3][0-9])- ]]; then
        if [[ ${BASH_REMATCH[1]}-${BASH_REMATCH[2]: -2}-${BASH_REMATCH[3]: -2} > $refdate ]]; then
            echo "$base"
        fi
    fi
done

Output:

202201013-020124
202201014-020123

• The regex ^([0-9]{4})(0[1-9]|01[0-2])(0[1-9]|0[1-3][0-9])- will match the formatted date, assigning the bash array BASH_REMATCH to the matched year, month, and day.
• The expression ${BASH_REMATCH[1]}-${BASH_REMATCH[2]: -2}-${BASH_REMATCH[3]: -2} reformats the array values into the date string as yyyy-mm-dd.
• Then the date string is compared with the reference date 2022-01-10.

If the month is between October and December and the string is formatted as 202201001 (1st of October, 2022), it will also work.

CodePudding user response：

A much simpler approach is to use awk. Simply split the last field of each record on '-' and compare against 20220110:

$ ls -1 /rec/flux_entrant/le5/archive/tble91_formation/* |
awk -F'/' '{split($NF, a, "-"); if (a[1]>20220110) print $NF}'

(note: above is ls -(one) not ls -(el))

You can tailor the search for files further by using find instead of ls. Up to you.

Result:

202201013-020124
202201014-020123

You can further sort or split as needed.