Announcement

**jlittle** · Feb 22, 2013, 07:40 PM

Originally posted by farjibaba View Post

... make this in bash...

(Well, bash can do this stuff, but in the not very long run, I'd suggest learning perl or python.)

The fourth hit in google for "date handling in bash" gave me this function:

Code:

date2stamp () {
    date --utc --date "$1" +%s
}

Unfortunately, date doesn't grok your dates, so a little fix up is needed:

Code:

date2stamp () {
    d1=${1//\// } # remove slashes
    d2=${d1/:/ } # remove first colon
    date --utc --date "$d2" +%s
}

This gets the dates into a form that can be compared. Assuming your user can enter the dates and the file on the command line:

Code:

(( $# < 3 )) && echo Usage: ${0##*/} start-date end-date file && exit  44

start=$(date2stamp $1) || exit 45
  end=$(date2stamp $2) || exit 46

while read date data; do
    date_utc=$(date2stamp $date)
    (( date_utc >= start && date_utc <= end )) && echo $date $data
done < $3

(44, 45, 46 are arbitrary).

Regards, John Little

**jlittle** · Feb 24, 2013, 10:53 PM

I started to feel, well, if not guilty, perhaps disappointed, in presenting such an inefficient solution as an example to a learner. The techniques used are typical for bash scripts but invoking a program (date) for every line of input is always going to be slow, as is invoking a bash function using $(). So here's a pure bash approach.

Code:

#!/bin/bash
declare -A mn=([Jan]=01
               [Feb]=02
               [Mar]=03
               [Apr]=04
               [May]=05
               [Jun]=06
               [Jul]=07
               [Aug]=08
               [Sep]=09
               [Oct]=10
               [Nov]=11
               [Dec]=12)

function comparable_date {
    (( ${#1} < 11 )) && { echo date:time '"'$1'"' too short; return 42; }
    local day=${1:0:2} month=${1:3:3} year=${1:7:4} hms=${1:12}
    local nm=${mn[$month]}
    [[ $nm == "" ]] && { echo month '"'$month'"' not understood; return 44; }
    eval $2=\"$year-$nm-$day-$hms\"
}

(( $# < 3 )) && { echo Usage: ${0##*/} start-date end-date file; exit 45; }

comparable_date $1 start || exit 46
comparable_date $2 end   || exit 47

while read date data; do
    comparable_date $date this &&
        [[ $this < $start || $this > $end ]] || echo $date $data
done < $3

Runs about 20 times faster on my computer. Note the tricky logic comparing the dates is because bash doesn't have ">=" or "<=" for strings.

Regards, John Little

Announcement

script for fetching data from a file

script for fetching data from a file

Comment

Comment