Reputation: 27
Here's my file:
$ cat myfile.txt
48700039|09:39:58|09:40:34
48700040|09:59:12|09:59:42
48700041|10:01:08|10:05:47
48700042|10:50:53|10:51:24
I want to subtract column 2 from 3, so my desired output would be:
48700039|09:39:58|09:40:34|00:00:36
48700040|09:59:12|09:59:42|00:00:30
48700041|10:01:08|10:05:47|00:04:39
48700042|10:50:53|10:51:24|00:00:31
I've tried everything... thanks in advance! =)
Upvotes: 1
Views: 1523
Reputation: 4496
Using sed and date,
while read line;do
#extraction of the columns 2 and 3 && converting them into an array
d=($(echo $line | sed -e 's/\(.*\)|\(.*\)|\(.*\).*/\2\n\3/'))
#substraction of dates
datediff=$(($(date -d ${d[0]} +%s)-$(date -d ${d[1]} +%s)))
#absolute value
datediff=${datediff/-/}
#print the line and the 'extra' H:M:S
printf "%s|%02d:%02d:%02d\n" $line $((datediff/3600)) $((datediff%3600/60)) $((datediff%60))
done < myfile.txt
Output:
48700039|09:39:58|09:40:34|00:00:36
48700040|09:59:12|09:59:42|00:00:30
48700041|10:01:08|10:05:47|00:04:39
48700042|10:50:53|10:51:24|00:00:31
Upvotes: 0
Reputation: 784998
GNU awk (gawk) based solution:
gawk -F"|" 'BEGIN {DT="2000 01 01 "} { gsub(/:/, " "); t1=mktime(DT $2);
t2=mktime(DT $3); gsub(/ /, ":"); print $0"|"strftime("%T", t2-t1, "UTC") }'
Upvotes: 0
Reputation: 45644
Sometimes it is necessary to leave the wonderful world of awk/sed/bash and move into a scripting language that understand times. This is such an occasion.
#!/usr/bin/env python
from datetime import datetime
FMT = '%H:%M:%S'
with open("myfile.txt") as fd:
for line in fd:
line = line.strip()
t = line.split('|')
tdelta = datetime.strptime(t[2], FMT) - \
datetime.strptime(t[1], FMT)
print "%s|%s" % (line, tdelta)
output:
48700039|09:39:58|09:40:34|0:00:36
48700040|09:59:12|09:59:42|0:00:30
48700041|10:01:08|10:05:47|0:04:39
48700042|10:50:53|10:51:24|0:00:31
Upvotes: 1