In a shell script I would like to calculate the difference between two timestamps.
It should look like similar to:

mystart=`date`
....
do something
....
myend=`date`
elapsed=calcdiff($mystart, $myend)
echo elapsed time=$(elapsed +%%d:%%H:%%M:%%S)

However the script above does not work.

Further question when the runtime is greater than 31 days does %%d above then contain e.g. "35"
or does it contain a value modulo 31 (here: 4) ?

Matthew

In the crontab I found some repetition values with a slash like

*/5

or

*/4

What does that mean e.g. in the second (=hpur) column?

Every 4 hours? Or 4 times per hour?

Wolfgang

From the command line I called (with nohup) a shell script which contains empty lines
(for better readability).

The nohup executes the script successfully but outputs a warning for the empty line:

/usr/local/joch/bin/dotest: line 3:
: command not found

How can I suppress this warning (of cause without removing the empty line)?

Do I really have to always avoid empty lines in shell scripts?

Jochen

Hi,
I have problem with lynx dump output. I have following dump structure
generated from lynx -dump.

Page
[1]First link name
[2]Second link name
..
..
..
[N]Nth link name

References

1. http://link_to_first
2. http://link_to_second
...
N. http://link_to_nth

Hi all!

I was running the following line on the UNIX terminal of my old G4 with
the latest version of Mac OS Tiger with no problem at all:

cut -f1-1 -d, fullx > ty1 ; cut -f2-2 -d, fullx > ty2 ; cut -f3-3 -d,
fullx | tr A-Z a-z > ty3

But now with THE SAME FILE, my new Mac pro with the latest Leopard, I
get the following error message:

cut: fullx: Illegal byte sequence
cut: fullx: Illegal byte sequence
cut: fullx: Illegal byte sequence

I noticed the message only occurred when a line with accented characters
(o umlaut or e acute accent) are encountered. If I replace those
characters with basic ASCII the process go as smoothly as with my old Tiger.

Is there an extra precaution I should take with my "cut" function or
some treatment I should apply to the file "fullx" for the ASCII
xenophobia to end?

Thank you!

First off, I should mention that I am a sed neophyte. What little I
know I've cobbled together over the years, and each new use of it
feels like learning it over from scratch. That being said, I love the
power of it.

So, I'm trying to write a bash script that will do a replace among
several files; I've got the rest of the script down, so I simplified
it for this question (nevermind that the html doesn't work). When I
run this script, it either errors out with an "unterminated
substitution" error, or it just does nothing... Here's some code:

#!/bin/bash
NUM=4
ITEM=A001
sed s#'\<\!-- Thumb$NUM --\>'#' \\'#g
A001.html > new_A001.html

the line "" (Thumb2,3,4, etc) is in each template file
that I am running this command on...

I have the following script. I am testing logical and of the return
values of the functions f and g. But I am getting the error. I am
calling the function g with two parameters t and y. Can you please
tell what I did wrong.
Thanks. Pedro

# cat funct.sh
#!/usr/bin/ksh

function f {
return 0
}

function g {
Var1=$1
Var2=$2
return 0
}

if f && "g t y"
then
echo done
fi

#!/bin/sh

awk '{
bo = substr($0,13,3)
slm = substr($0,150,8)
slo = substr($0,175,7)
inc = substr($0,97,10)/100
busi = substr($0,83,10)
mth = substr($0,39,2)
yer = substr($0,35,4)
printf"%%3s:%%8s:%%7d:%%10.2f:%%10d:%%.2d:%%4d
\n",bo,slm,slo,inc,busi,mth,yer
}' infile | sort -t: +6 -7 +5 -6 +2 -3n +1 -2d > tmpout

I have a file with several text blocks, for example a file with these
2 blocks, and I want to sort the file such that each block is sorted
(with a sort -rn):

33.2 blah
57.72 foo
100.0 bar

23.3 sdf
64.76 jzsdt

So the output should be:

100.0 bar
57.72 foo
33.2 blah

64.76 jzsdt
23.3 sdf

I have no idea how to do this on the shell. I can do it in perl, but
there the numerial sort function sort {$a <=> $b} complains if there
are also lexical elements on a line, and a shell solution might be
faster anyway. The files I want to run this on typically have several
hundreds of these blocks.

Hi,

I am trying to write a shell program to do the same task on different
files. My file names differ by number, something like

chr1.in
chr2.in
chr3.in ... etc.

How can use, awk '{$1 = 1; print $1"\t"$2"\t"$3"\t"$4}' chr1.in >
chr1.out for all the files? I tried with sprintf in a for loop but
didn't work.

Thanks in advance.

Regards,
Ezhil