I am trying to work out a Linux command to split a large log file into pieces based on date.

Using How to split existing apache logfile by month? as a starting point, I tried:

awk '{ split($4,array,"/"); print > array[2] ".txt" }' TestLog.txt

On my sample TestLog.txt with entries for May, Jun, and Jul of different years, this created text files May.txt, Jun.txt and Jul.txt:

In order to understand the values in the arrays, I eliminated the output file, and displayed the array values using:

awk '{ split($4,array,"/"); print  array[1] "  "  array[2] "  " array[3] "  " array[4] }' TestLog.txt

Where the first 2 lines of TestLog.txt are: - - [30/May/2011:23:21:37 -0500] "GET / HTTP/1.0" 200 206492 "-" "Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; SV1; .NET CLR 1.1.4322;TencentTraveler)" - - [31/May/2011:00:36:35 -0500] "GET / HTTP/1.0" 200 206492 "-" "Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; SV1; .NET CLR 1.1.4322;TencentTraveler)"

This resulted in [30 May 2011:23:21:37 for the first line in the file.

The results were very confusing to me. In particular:

  1. Why is array[1] equal to [30 and not - - [30 ?

  2. Why is array[3] equal to 2011:23:21:37 and not 2011:00:36:35 -0500] "GET?

  3. Why is array[4] null?

  4. What should the value of array[0] be?

  • Why did you choose to split $4? Apr 5, 2016 at 2:14
  • I just used what had been specified at the link I cited. I assumed it was specifying that the line should be split into 4 pieces. Based on the explanation below, I now understand that $4 gabbed the 4th string ([30/May/2011:23:21:37) for further spitting.
    – Mike
    Apr 5, 2016 at 2:45

1 Answer 1


Lets take the first line: - - [30/May/2011:23:21:37 -0500] "GET / HTTP/1.0" 200 206492 "-" "Mozilla/4.0 (compatible; MSIE 6.0; Windows NT 5.1; SV1; .NET CLR 1.1.4322;TencentTraveler)"

and the crucial part of the awk snippet:

awk '{ split($4,array,"/") ...

Here what is happening:

  • awk runs and splits the line on the spaces (default field separator)
  • 4th field in the line is additionally being split on / character
  • the result of the split is put into the array
  • later on the whole line is printed to the file named as a second subfield (array[2]) of the 4th field

so $4 field initially contained [30/May/2011:23:21:37, and after split we have


There is no array[4], because there the 4th field doesn't contain 4th "subfield" and there is no array[0] because in awk array indexes start from 1.

  • 1
    Aha, I couldn't understand what he was trying to do. So in this case he could just use awk -F/ '{ print > $2 ".txt" }' TestLog.txt.
    – Wildcard
    Apr 5, 2016 at 0:22
  • @Wildcard very likely, if it is guarantee that there are no / before date (judging from the link to SO thats the case).
    – jimmij
    Apr 5, 2016 at 0:27
  • I wondered if spit had a default delimiter of space. So those splits are basically being ignored, and the 4th string, (delimitered by spaces) is then split using / as the delimiter. Then how can I extract the year (2011) also?
    – Mike
    Apr 5, 2016 at 2:39
  • @Mike You can add : as a field separator: awk -F '[/:]' '{print "Month: "$2", year: "$3}'
    – jimmij
    Apr 5, 2016 at 8:40

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .