UNIX for Dummies Questions & Answers

There are files in a directory and I have to move multiple files adding datetimestamp before the file type.

Code:

/Data/

abc.csv
def.csv
ghi.csv

I have to move this files to archive directory adding datatimestamp before .csv

/archive/
abc_YYYYMMDDHHMMSS.csv
def_YYYYMMDDHHMMSS.csv
ghi_YYYYMMDDHHMMSS.csv

Works with GNU date on modern linux bash:

Code:

#! /bin/bash

dt=$(date +%Y%m%d%H%M%S)
for file in /Data/*.csv
do
    name=${file%.*}
    mv /Data/$file /archive/${name}_${dt}.csv
done

is there any sed or awk command that does this logic. The reason being I work on an ETL tool which mainly support one line logic otherwise I have to maintain a shell script beyond the tool.

Not quite. The first time through this loop with the sample data specified by the OP would try to move /Data/Data/abc.csv to /archive/Data/abc_timestamp.csv.

I would tend to parameterize it a little bit more with something like:

Code:

#!/bin/ksh
destdir="${2:-/archive}"
srcdir="${1:-/Data}"
if [ "${destdir:0:1}" != "/" ]
then    destdir="$PWD/$destdir"
fi
if [ ! -d "$srcdir" ] || [ ! -d "$destdir" ]
then    printf "Source (%s) and target(%s) directories must exist.\n" \
                "$srcdir" "$destdir" >&2
        printf "Usage: %s source_directory target_directory\n" "${0##*/}" >&2
        exit 1
fi
cd "$srcdir"
ts=$(date +%Y%m%d%H%M%S)
for file in *.csv
do      mv "$file" "$destdir/${file%.*}_$ts.csv"
done

I usually use the Korn shell, but this will work with recent versions of ksh and bash without change. It defaults to /Data and /archive as the source and target directories but they can be overridden by supplying (absolute or relative) directory names as command line arguments. It also protects itself in case any of the components of the source and target files contain whitespace characters by quoting user supplied data. Production code should also verify that the cd succeeds and that all of the mv commands succeed.

But, now that I see that the OP has changed the requirements, this is obviously overkill. You could create a shell script using sed and write it as a 1-liner, write an awk script that invokes the shell to get the date and perform the moves, or convert a corrected shell script to a 1-liner but I won't help you play that game. I will always go for a readable, more easily maintainable script than a hard to read 1-liner.

I don't think you can dream up a sed or awk one liner as you need file system interaction to move the files; perl might do. Did you consider the rename perl script?

No, I am not so good at unix scripting.

Can one line works with moving multiple files adding date at the end?

Code:

abc.txtYYYYMMDDHHMMSS

You can do most things with a one line shell script that you can do with a 500 line shell script except edit it easily, understand it, debug it, and make changes to it that are likely to work the first time (or ever).