Home / ubuntu / Questions / 477779
Accepted
Roboman1723
Asked: 2014-06-07 10:14:28 +0800 CST

Document Sorting Section

  • 772

I have some sample data below (call it test.txt). I would like to split this big document into 3 documents by searching the first two digits in the 4th column. I'm pretty new with Linux (Lubuntu 14.04) and have looked into 'csplit' and 'awk' but can't seem to get the syntax down for either. (I'm writing BASH scripts)

Before:

test.txt

12-1-2014   Allow   00:00:00 00:00:00
12-1-2014   Allow   00:00:00 00:00:00   
12-2-2014   Allow   01:00:00 01:00:00
12-10-2014  Deny    01:00:00 01:00:00
12-10-2014  Deny    02:00:00 02:00:00
12-11-2014  Deny    02:00:00 02:00:00

After:

test1.txt

12-1-2014   Allow   00:00:00 00:00:00
12-1-2014   Allow   00:00:00 00:00:00

test2.txt

12-2-2014   Allow   01:00:00 01:00:00
12-10-2014  Deny    01:00:00 01:00:00

test3.txt

12-10-2014  Deny    02:00:00 02:00:00
12-11-2014  Deny    02:00:00 02:00:00
lubuntu

3 Answers

  1. Best Answer

    Something like that maybe : 

    #!/bin/bash

while read LINE ; do
 [[ -n "`echo $LINE | awk '{print " "$NF}' | grep ' 00:'`" ]] && echo "$LINE" | tee -a test1.txt
 [[ -n "`echo $LINE | awk '{print " "$NF}' | grep ' 01:'`" ]] && echo "$LINE" | tee -a test2.txt
 [[ -n "`echo $LINE | awk '{print " "$NF}' | grep ' 02:'`" ]] && echo "$LINE" | tee -a test3.txt
done < test.txt

    Where : 

    while read LINE ; do ; done < FILE

    is a loop that reads every line of FILE

    [[ -n "`<command>`" ]] &&

    means : if <command> outputs a non-zero lenght string, then...

    echo "these are something"| awk {print $NF} | grep 'thing'

    means : read "these are something", keep only the last column, and see if "thing" is in it.

    echo "something" | tee -a FILE

    means : I print "something" in FILE, without overwriting anything and by creating it if it doesn't exist yet.

    • 0

  2. With awk, you could just do

    awk '{split($3,a,":"); print > "test"a[1]+1".txt";}' test.txt


    In pure bash I'm not sure the best way - although you could do something like

    while read -r line; do 
  read -r a b c d <<< "$line"
  printf -v outfile "test%d.txt" $(( ${c%%:*} + 1 ))
  printf "%s\n" "$line" >> "$outfile"
done < test.txt
    • 0

  3. Try this script:

    #!/bin/bash
while read line; do 
HOUR=`echo $line | cut -d' ' -f3 | cut -d':' -f1`;
echo $line >> test-$HOUR.txt;
done < test.txt

    For each line in file, command cut splits line first for space, then for :. The result is used as suffix for file.

    Output are:

    file-00.txt  file-01.txt  file-02.txt

    that containing line for hour 00, 01 and so on.

    • 0