find command generates empty files in linux

Question

I am trying to retrieve the filesize associated with the Logfilename listed in ip_DBfiles.txt with the logfilenames from the server.When matched the size of the file and the filename should be written to 3rd file op_DBfiles.txt

1.ip_DBfiles.txt

Date          Logfilename
01/06/2021 /server/base/a.log
02/06/2021 /server1/base1/b.log

2.cd to /server/lgs

db_listfiles="$(cat ${ip_DBfiles} |awk '{print $3}')"

for i in "${db_listfiles[@]}"
  do

   find . -type f -print|egrep "(${i})" > /dev/null 2>&1
   if [ $? -eq 1 ] ;
         then
             echo "no action"
         else
            echo
            find . -type f -print|egrep "(${i})" -exec du -ah {} \; > filesize.txt                            
    fi
   done

op_DBfiles.txt

Number  Date          Logfilename      size
1      01/06/2021 /server/base/a.log     5
2      02/06/2021 /server1/base1/b.log    6

when using stat I get the following error msg

stat: cannot stat `/server/base/a.log\n/server1/base1/b.log': No such file or directory

Though the file exists it generates empty file in filesize.txt with find command

Appreciate your help!

What is the difference between command prompt and bash in your case (what do you mean)? — sudodus
– sudodus, Commented Nov 28, 2021 at 2:54
What happens if you add a shebang line at the top of the file? #!/bin/bash; It is possible that different shells are executing your script file, and the shebang line will make sure that bash will execute it (if that is what you want). - Also, the input ${db_files[@]}, are you sure that it is found in both cases? — sudodus
– sudodus, Commented Nov 28, 2021 at 3:03
-name "(${i})" will only match names exactly matching (${i}) (as a shell glob), whereas -print|egrep "(${i})" will match any files whose name contains (${i}) (as an extended regular expression). You haven't shared enough information to know whether that's significant. — steeldriver
– steeldriver, Commented Nov 28, 2021 at 3:30
edit your question to include a complete (but minimal) script that shows the problem, and also show how you run the script. And also what output you expect to get, and in which file. Given the script you posted doesn't set b_file at all, I'm guessing it's not the full script — ilkkachu
– ilkkachu, Commented Nov 28, 2021 at 11:54
It looks like you're setting db_listfiles as a plain variable (with filenames separated by newlines), but trying to use it as if it were an array. Probably the simplest way to fix this would be to use it without double-quotes (for i in ${db_listfiles}) and hope that word-splitting splits it correctly. Also, > filesize.txt will delete all previous contents each time through the loop; you probably want >> filesize.txt. BTW, I had to fix your formatting. See this reference. — Gordon Davisson
– Gordon Davisson, Commented Nov 29, 2021 at 0:27

cas · Accepted Answer · 2021-11-28 05:06:31Z

You're running Linux, so you can use GNU's version of stat (which is in the GNU coreutils package), which has a very useful --printf option for formatting the output. You don't need the for loop at all.

Assuming that the "${db_files[@]}" array contains the full paths to the filenames (or, at least, pathnames relative to the current directory), you can run something like this to get the file sizes (in bytes) and names:

stat --printf "%s\t%n\n" "${db_files[@]}"

This prints the file size, a tab, the filename, and then a newline for each filename in db_files.

If you want the sizes in "human-readable" format, you can use numfmt (also in GNU coreutils) to reformat them:

stat --printf "%s\t%n\n" "${db_files[@]}" | numfmt --to=si

e.g.

$ stat --printf "%s\t%n\n" ./file*.txt 
12942   ./file1.txt
71529   ./file2.txt
83135   ./file3.txt
12889   ./file4.txt


$ stat --printf "%s\t%n\n" ./file*.txt | numfmt --to=si
13K ./file1.txt
72K ./file2.txt
84K ./file3.txt
13K ./file4.txt

BTW, if you want to use a loop, you could do it like this:

for f in "${db_files[@]}"; do
  if [ -f "$f" ] ; then
    stat --printf "%s\t%n\n" "$f" | numfmt --to=si >> "$b_file"
  else
    echo "$f" >> "$a_file"
  fi
done

Stack Exchange Network

find command generates empty files in linux

1 Answer 1

You must log in to answer this question.

Hot Network Questions

find command generates empty files in linux

1 Answer 1

You must log in to answer this question.

Related

Hot Network Questions