Replacing a parameter in a text file based on the latest filename in a directory - bash

Question

I have the following text file (namelist.txt)
&share
wrf_core = 'ARW',
max_dom = 3,
start_date ='YYYY-MM-DD_HH:00:00','YYYY-MM-DD_HH:00:00','YYYY-MM-DD_HH:00:00',
end_date ='YYYY-MM-DD_HH:00:00','YYYY-MM-DD_HH:00:00','YYYY-MM-DD_HH:00:00',
interval_seconds = 21600,
io_form_geogrid = 2,
debug_level=0,
/
I want to replace the YYYY, MM, DD, and HH based on the latest filename of a directory.
For example:
An INPUT folder contains the following subdirectories:
2021021000
2021021006
2021021012
2021021018
2021021100
The latest directory from the above is 2021021100
I'm stuck here. The script should read the latest filename of the sub-directory inside the INPUT folder and do the following.
year=$(echo $line | cut -c1-4)
echo $year
month=$(echo $line | cut -c5-6)
echo $month
day=$(echo $line | cut -c7-8)
echo $day
hour=$echo $line | cut -c9-10)
echo $hour
sed -i 's/'YYYY'/'$year'/g' namelist.txt
sed -i 's/'MM'/'$month'/g' namelist.txt
sed -i 's/'DD'/'$day'/g' namelist.txt
sed -i 's/'HH'/'$hour'/g' namelist.txt
The desired output should be like this:
&share
wrf_core = 'ARW',
max_dom = 3,
start_date ='2021-02-11_00:00:00','2021-02-11_00:00:00','2021-02-11_00:00:00',
end_date ='2021-02-11_00:00:00','2021-02-11_00:00:00','2021-02-11_00:00:00',
interval_seconds = 21600,
io_form_geogrid = 2,
debug_level=0,
/
How can I do this in bash?
I'll appreciate any help on this.

Answer 1

Get the directory with the latest date
Bash's globs (* and so on) expand in a sorted order. If the subdirectories in your current working directory are only named in the style YYYYMMDDHH then */ will expand to a list of dates where the last date is at the end of the list. To retrieve only the last entry from that list you can use either an array, a function (using shift), or a command (for instance printf | tail). Here we go with the array:
#! /bin/bash
cd INPUT
dirs=(*/)
last="${dirs[-1]}"
cd -
If there are also other directories you can change the glob so that only directories of the format YYYYMMDDHH are accepted:
dirs=([0-9][0-9][0-9][0-9][0-1][0-9][0-3][0-9][0-2][0-9]/)
Replacing the placeholders
You don't need four cut and four sed. The following should work as well:
sed -i "s/YYYY-MM-DD_HH/${last:0:4}-${last:4:2}-${last:6:2}_${last:8:2}/g" yourFile

Answer 2

GNU Awk is a possibility for this:
awk -v dat="2021021100" 'BEGIN { yr=substr(dat,1,4);mn=substr(dat,5,2);day=substr(dat,7,2);hr=substr(dat,9,2)} /YYYY-MM-DD_HH/ { gsub("YYYY-MM-DD_HH",yr"-"mn"-"day"_"hr,$0) }1' namelist.txt > namelist.tmp && mv -f namedlist.tmp namedlist.txt
Explanation:
awk -v dat="2021021100" # Pass the date as a variable dat to awk
'BEGIN {
yr=substr(dat,1,4); # Before processing the file, use substr to extract the time elements from dat
mn=substr(dat,5,2);
day=substr(dat,7,2);
hr=substr(dat,9,2)}
/YYYY-MM-DD_HH/ {
gsub("YYYY-MM-DD_HH",yr"-"mn"-"day"_"hr,$0) # When we find YYYY-MM-DD_HH" in the line, use gsub to substitute this for the yr,mn,day and hr.
}1' namedfile # Print all lines amended or otherwise
If you have more recent versions of GNU awk, can use the -i flag for "in file" changes as opposed to using a tmp file and actioning mv.