I'm new to scripting in bash and our professor had us sort a file in this format
peas|10.00|05 Apr 2012
pea soup|10.00|05 Jan 2012
ham|10.00|06 Apr 2012
using the date on the third field, with the most recent item appearing on top. I've tried using combinations of filters and sort but they didn't work. Can anyone help me? thanks
try
sort -t '|' -k 3.8,3.11nr -k 3.4,3.6Mr -k 3.1,3.2nr < input
------ ------------- ------------ ------------
sep first key second key third key
$ cat input.txt | awk -F '|' '{sprintf("date +%%s -d \"%s\"", $3) | getline tm}; {print tm "\t" $0}' | sort | cut -f2-
pea soup|10.00|05 Jan 2012
peas|10.00|05 Apr 2012
ham|10.00|06 Apr 2012
If you don't want call external command date,
you can write a custom mktime2 function in awk:
#!/bin/gawk -f
# script.awk
BEGIN {
FS="|"
m["Jan"] = "01"
m["Feb"] = "02"
m["Mar"] = "03"
m["Apr"] = "04"
m["May"] = "05"
m["Jun"] = "06"
m["Jul"] = "07"
m["Aug"] = "08"
m["Sep"] = "09"
m["Oct"] = "10"
m["Nov"] = "11"
m["Dec"] = "12"
}
{
print mktime2($3) "\t" $0 | "sort | cut -f2-"
}
function mktime2(s, arr,yyyy,mm,dd)
{
split(s, arr, " ")
yyyy = arr[3]
mm = m[arr[2]]
dd = arr[1]
return mktime(sprintf("%s %s %s 00 00 00", yyyy, mm, dd))
}
# make script executable
$ chmod +x script.awk
# run the script
$ ./script.awk input.txt
pea soup|10.00|05 Jan 2012
peas|10.00|05 Apr 2012
ham|10.00|06 Apr 2012
This may work for you (GNU sort):
sort -t'|' -k3.8,3.11nr -k3.4,3.6Mr -k3.1,3.2nr file
Or (if you don't have GNU sort):
sed '1{x;s/^/Jan01Feb02Mar03Apr04May05Jun06Jul07Aug08Sep09Oct10Nov11Dec12/;x};G;s/\(.*|\(..\) \(...\) \(....\)\)\n.*\3\(..\).*/\4\5\2 \1/' file
sort -nr |
sed 's/^[^ ]* //'
Similar to kev's answer, here's one that doesn't use awk
while IFS=\| read -r item price date ; do printf '%s|%s|%s|%s\n' "$(date +%s -d "$date")" "$item" "$price" "$date" ; done < table.txt | sort -n -t\| | cut -d\| -f2-
The idea is to add a field sort can use, sort by it, then strip it.
