Search Results

Do not use echo -e, do not use echo -n.
printf "%s\n" "mystring"

Minor re-write...
OSX 10.7.5, default bash terminal:-
#!/bin/bash
# nl_date.sh
> /tmp/txt
DATE=$1
myview()
{
base_name=$1
echo "$base_name$DATE" >> /tmp/txt
}
myview sample3030
myview...

Change the echo to print...

Try:-
echo -e "${base_name}${DATE}">> /path/to/file/txt

I guess it is extracting, but it will extract it to the current working directory where you are running your script.

You have to specify the directory:
tar -zxvf /path/to/the/file/file.tar.gz -C...

Try:-
[ ! -f /path/to/file ]

No, it should be if [ ! -f /path/to/file ]

Depending on the shell you use, which you fail to mention, you could also tryfor file in 2014-02-01_{aaaa,bbbb,cccc,dddd}.txt;

It searches the files in the current directory.
Either change to the desired directory first
cd /path/to/files || exitOr put the path into a variable and prefix it every time:
fpath=/path/to/files...

It overwrites the file. But you get the output of the entire loop in the file -- it gets opened once and written to lots of times.

Pass filename as input to sed:
for f in path/to/files/date_*.csv
do
sed '1d' "$f"
done > merge.csv

#!/bin/bash

mysub(){
if [ $1 -eq 0 ]
then
echo "NO CHANGES ARE NEEDED"
exit
else
sed -i '1i'"$2" $3
echo "ADDED HEADER"
fi
)

zcat 2013-12-26_file1.csv.gz >...

Read man test or check your shell's manual page (ksh and bash, for example, have a builtin version that they document).

Regards,
Alister

If you actually have multiple formats like:

yyyy/mm/dd
Wed 9 Sep 2009

And so on, then cleaning them up can be a nightmare. Try reading the tsv into Excel, and then have Excel produce a fixed...

So long as you have some consistency in the date format something like the following (adapted to your set of date string types)

#!/usr/bin/perl

use strict;
use warnings;

my %MON=qw(JAN 1...

Searching for password in man unzip:

-P password
use password to decrypt encrypted zipfile entries (if any).
THIS IS INSECURE! Many multi-user ...

add a block like the following to the start of the run to keep track of the counter in a file.

$next=$(( $(cat .last_run) + 1 ))
echo $next >> .last_run

Use substring extraction:
mkdir ${INPUTDATE:2:2} ${INPUTDATE:4:2} ${INPUTDATE:6:2}
Syntax:
${string:position:length}

I'm not sure what the hadoop command is doing. Anyway, maybe this might work:

SOURCE=/path/to/files/
find "$SOURCE" -type f -mtime -1 -exec cp -v "{}" /usr/path/ \;

Your program has already quit when you see that message -- it just looks odd, because nohup has printed junk overtop of your prompt right after.

Normally this doesn't happen because the shell...

awk '
/10000/ { # Search if pattern: 10000 present in current record
sub(/\/.*/, x, $NF) # If found, substitute every character...

var=$( netstat -plen | awk '/10000/{sub(/\/.*/,x,$NF);print $NF}' )

find /path/to/logs/ -mtime +10 ! \( -name '*.info' -o -name '*.extn' -o .... \) -exec echo '{}' ';'

Use https://www.unix.com/images/editor/code.png tags for code, please.

find is already recursive -- looks inside folders -- so you don't need to put it in a loop.

It will also take a -name...

Use logical not ! to exclude such files:
! -name "*.info"