FileList and sed

[RENOO]

Hello,

First time on this forum. Hope to be back here soon...

I'am doing simple scripts in bash. I use bash because it is the only language I kind of know.

I have 2 directories,called /input and /output. The first one /input is régularly updated with new files from an ftp. I need to copy the file from /input to /output if a value, the $ValueF, is smaller in the newer file. In order to do this I've created a text file called FileDatabase.txt in /output directory where I store the filename and the corresponding ValueF. My script check the ValueF in the FileDatabase.txt and then copy in /output if needed. My problem is that for some reasons I have hard time updating the ValueF in FileDatabase.txt. I'am using a sed command. But this command only works when I am pasting it in the terminal but not inside my script
Really !

Here is my script I have simplify for clarity.

Code: Select all

#!/bin/bash

# List all the files in INPUT directory. At each file correspond an ValueF. If the Value F is smaller
# in the INPUT directory, cp file from input directory to OUTPUT directory

partern=pattern
#Directories parameters
input_dir=/home/Data/INPUT
output_dir=/home/Data/OUTPUT


# On va d'abord looper sur tout les fichiers.
find ${input_dir} -name "$pattern" | sort -r > ${input_dir}/Input_filelist.txt
while read input_file;
do
  # Extract filename
  basefilename=`basename "$input_file"`
  echo "$basefilename"
  #Extract ValueF
  Input_ValueF=`grep ValueF "$basefilename" | cut -d"=" -f2`
 
  # remove leading 0
  ((Input_ValueF=10#${Input_ValueF}))

  # Output file (same name as input file, different directory
  output_file=${output_dir}/$basefilename
 
  if [ ! -f ${output_file} ]; then
    # First time, just copy to output dir
    echo "First Time"
    cp ${input_file} ${output_file}
    # File up database
    echo "${output_file} ${Input_ValueF}" >> ${output_dir}/FileDatabase.txt
    sort -o ${output_dir}/FileDatabase.esv ${output_dir}/FileDatabase.txt
  else
    echo "Check FValue"
    # Output file already exists.
    # Check ValueF reading FileDatabase.txt
    ValueF=`grep ${output_file} ${output_dir}/FileDatabase.txt | cut -d" " -f2`

    echo "$Input_ValueF vs $ValueF "
   # Compare both Value F
    if [[ "$Input_ValueF" -lt "$ValueF" ]]; then
      echo "better one"
      # Overwrite old file with new one
      cp ${input_file} ${output_file}
      # Update database using sed command
      sed  '/${output_file}/ s/ ${ValueF} / ${ValueF} /' ${output_dir}/FileDatabase.txt > new
      #wait ${!}
      mv new ${output_dir}/FileDatabase.esv
    fi
  fi
  echo
done < ${input_dir}/Input_filelist.txt


What is wrong with my sed command ??
If you have a better way and more reliable way to do this kind of thing with file, I'll be pleased to learn. ( shall I use a database ? how does it work ? )

THanks in advance

RENOO

[Watael]

the problem comes from sed's separator, change to some character that is not in file's path (something like @)

[RENOO]

Hi,

Thank you very much for your reply.
You've mentionned a very good point here. I thought we had it.

It did the following switch

Code: Select all

sed  '/${output_file}/ s@ ${ValueF} @ ${Input_ValueF} @' ${output_dir}/FileDatabase.txt > new



As modifying the 2 firsts "/" gives an error.
Unfortunatly, it is still not working. The "new" file is similar than the FileDatabase.txt ??

What am I doing wrong ?

[Patsie]

Code: Select all

sed  '/${output_file}/ s@ ${ValueF} @ ${Input_ValueF} @' ${output_dir}/FileDatabase.txt > new

In your sed there are spaces before and after both ${ValueF} and ${Input_ValueF}. Are there spaces there in the FileDatabase.txt file (especially the spaces behind the values).

[Watael]

Code: Select all

sed  '\@${output_file}@s@${ValueF}@${Input_ValueF}@' ${output_dir}/FileDatabase.txt > new

[RENOO]

Hi there,

I confess i have not been very smart here. I don't need to create the databased.txt as the FValue is contain inside the file itself. I just need to read the file itself instead of creating another file and modifying the FValue.
My mistake here. But still I very thank you for you help. With the time I am getting better...

@Watael
Why do you only "escape" (\) the first @ in your code ? How shall I understand this ?

Thanks again all of you.
RENOO

[Watael]

GNU Sed Manual:

3.5 The s Command
[...]The / characters may be uniformly replaced by any other single character within any given s command. The / character (or whatever other character is used in its stead) can appear in the regexp or replacement only if it is preceded by a \ character.