How to randomly select lines from a text file


 
Thread Tools Search this Thread
Top Forums UNIX for Dummies Questions & Answers How to randomly select lines from a text file
# 1  
Old 10-23-2012
How to randomly select lines from a text file

I have a text file with 1000 lines, I want to randomly select 200 lines from it and print them as output. How do I go about doing that? Thanks!
# 2  
Old 10-23-2012
Your time would be better spent learning awk than asking dozens of questions which could be trivially answered in it.

Code:
awk 'NR==FNR { B++; next }
(NR != FNR) && (!Z) {
        srand();
        if(B < 200) exit(1); # Too few lines
        for(N=1; N<=200; )
        {
                V=sprintf("%d", (rand()*B)+1)+0;
                if(!(V in A)) { A[V]=1; N++ }
        }
        Z=1
} NR in A' inputfile inputfile


Last edited by Corona688; 10-25-2012 at 12:34 PM.. Reason: Fixed typos
This User Gave Thanks to Corona688 For This Post:
# 3  
Old 10-23-2012
For the fun of it here's another way that does not use awk although the awk version will be more efficient. This has the overhead of creating the pipeline repeatedly which should be avoided for good practice. Also I believe the ksh RANDOM built-in has a limit of 32767 that must be considered if the file is large.
Code:
$ cat x
##
## x nbr_of_lines_wanted  filename
##
#!/bin/ksh

iterations=$1
file="$2"

((lines_avail=$(wc -l < "$file")+1))

while (( $iterations > 0 )); do
  head -$((${RANDOM} % $lines_avail)) "$file" | tail -1
  (( iterations=$iterations - 1 ))
done

exit 0

This is actually a good example of how a seemingly simple solution for a small file can end up burning you on performance and system limitations should you need to run it on a much larger file
or a system that may see increased load in the future.
Typically when you see a long command line or pipeline like this being done a large number of times (especially a user-enterable number of times) it should
be a red flag warning that there will most likely be a more efficient way of structuring the program.

Last edited by gary_w; 10-23-2012 at 05:11 PM..
This User Gave Thanks to gary_w For This Post:
# 4  
Old 10-24-2012
Hi.

There are a number of commonly-available utilities to do this. Here is a demonstration of two:
Code:
#!/usr/bin/env bash

# @(#) s1	Demonstrate random selection of lines with rl, shuf.

pe() { for _i;do printf "%s" "$_i";done; printf "\n"; }
pl() { pe;pe "-----" ;pe "$*"; }
db() { ( printf " db, ";for _i;do printf "%s" "$_i";done;printf "\n" ) >&2 ; }
db() { : ; }
C=$HOME/bin/context && [ -f $C ] && $C rl shuf

pl " Lines selected from:"
cat data0

# Prepare data file from single line of words.
tr ' ' '\n' < data0 > data1

pl " Results from shuf, 1:"
shuf -n 3 data1

pl " Results from shuf, 2:"
shuf -n 3 data1

pl " Results from rl, 1:"
rl -c 3 data1

pl " Results from rl, 2:"
rl -c 3 data1

exit 0

producing:
Code:
% ./s1

Environment: LC_ALL = C, LANG = C
(Versions displayed with local utility "version")
OS, ker|rel, machine: Linux, 2.6.26-2-amd64, x86_64
Distribution        : Debian GNU/Linux 5.0.8 (lenny) 
bash GNU bash 3.2.39
rl 0.2.7
shuf (GNU coreutils) 6.10

-----
 Lines selected from:
foo bar baz qux quux corge grault garble warg fred plugh xyzzy thud

-----
 Results from shuf, 1:
quux
thud
qux

-----
 Results from shuf, 2:
thud
xyzzy
quux

-----
 Results from rl, 1:
corge
qux
grault

-----
 Results from rl, 2:
thud
corge
warg

You may need to install these from your distribution repository. See man pages for details.

See also Algorithm::Numerical::Sample - search.cpan.org if a perl module is desirable.

Best wishes ... cheers, drl
This User Gave Thanks to drl For This Post:
# 5  
Old 10-24-2012
In the order of lines in the file, without all lines in memory:
Code:
awk '
  NR==FNR { next }
  FNR==1{
    srand;
    n=NR-1
    for(i=1; i<=200; i++) {
      line=0
      while(!line || line in A) line=int(rand*n)+1
      A[line]
    }
  } 
  FNR in A
' infile infile


In the order of the selection, with all lines in the file in memory..
Code:
awk '
  { R[NR]=$0 }
  END{
    srand;
    n=NR
    for(i=1; i<=200; i++) {
      line=0
      while(!line || line in A) line=int(rand*n)+1
      A[line]
      print R[line]
    }
  } 
' infile

# 6  
Old 10-24-2012
Code:
cat file | head -n 500 | tail -n 200

A very simple idea but not for random lines.
# 7  
Old 10-24-2012
Quote:
Originally Posted by Corona688
Code:
awk 'NR==FNR { B++; next }
(NR != FNR) && (!Z) {
        srand();
        if(Z < 200) exit(1); # Too few lines
        for(N=1; N<=200; N++)
        {
                V=sprintf("%d", (rand()*B)+1)+0;
                if(!(V in A)) { A[V]=1; N++ }
        }
        Z=1
} NR in A' inputfile inputfile

Looks like there are a couple of bugs there. The first, a typo of Z for B, always exits before selecting anything. The second, the for-loop post-increment expression, can cause the algorithm to yield anywhere between 1 and 200 lines.

Probably no need to fix it since Scrutinizer's first example in post #5 is a correct implementation of the same approach.

Regards,
Alister

---------- Post updated at 06:28 PM ---------- Previous update was at 06:26 PM ----------

With regard to all of the AWK suggestions, without knowing exactly how the script is to be used, it's possible that all of the recommendations are inadequate. Nearly every awk srand implementation's default seed is the number of seconds since the epoch. Successive or simultaneous runs could yield identical results. May or may not be an issue. We don't have sufficient information to make that determination. Just a head's up for the OP.

If it is an issue, more information would be required to determine a robust seed expression.

Regards,
Alister
This User Gave Thanks to alister For This Post:
 
Login or Register to Ask a Question

Previous Thread | Next Thread

10 More Discussions You Might Find Interesting

1. Shell Programming and Scripting

Match text to lines in a file, iterate backwards until text or text substring matches, print to file

hi all, trying this using shell/bash with sed/awk/grep I have two files, one containing one column, the other containing multiple columns (comma delimited). file1.txt abc12345 def12345 ghi54321 ... file2.txt abc1,text1,texta abc,text2,textb def123,text3,textc gh,text4,textd... (6 Replies)
Discussion started by: shogun1970
6 Replies

2. Shell Programming and Scripting

How to select lines randomly without replacement in UNIX?

Dear Folks I have one column of 15000 lines and want to select randomly 5000 of them in five different times without replacement. I am aware that command 'shuf' and 'sort -R' could select randomly those lines but I am not sure how could I avoid the replacement of selection line. Does anyone have... (10 Replies)
Discussion started by: sajmar
10 Replies

3. Shell Programming and Scripting

Select lines from a file based on a criteria

Hi I need to select lines from a txt file, I have got a line starting with ZMIO:MSISDN= and after a few line I have another line starting with 'MOBILE STATION ISDN NUMBER' and another one starting with 'VLR-ADDRESS' I need to copy these three lines as three different columns in a separate... (3 Replies)
Discussion started by: Tlcm sam
3 Replies

4. Shell Programming and Scripting

Get 20% of lines in File randomly

Hello, This is my code: nb_lignes=`wc -l $1 | cut -d " " -f1` for i in $(seq $nb_lignes) do m=`head $1 -n $i | tail -1` //command done Please how can i change it to get Get 20% of lines in File randomly to apply "command" on each line ? 20% or 40% or 60 % (it's a parameter) Thank you. (15 Replies)
Discussion started by: chercheur857
15 Replies

5. Shell Programming and Scripting

randomly shuffle two text files the same way

What I have are two text files that I need to shuffle randomly, but I need the two files to be randomly shuffled the same way. I have heard of shuf but I do not know how to use it for two files. Maybe there is also an easy/simple awk command I do not know about that could handle this problem. ... (3 Replies)
Discussion started by: adrunknarwhal
3 Replies

6. Shell Programming and Scripting

Select lines in which column have value greater than some percent of total file lines

i have a file in following format 1 32 3 4 6 4 4 45 1 45 4 61 54 66 4 5 65 51 56 65 1 12 32 85 now here the total number of lines are 8(they vary each time) Now i want to select only those lines in which the values... (6 Replies)
Discussion started by: vaibhavkorde
6 Replies

7. Shell Programming and Scripting

Select some lines from a txt file and create a new file with awk

Hi there, I have a text file with several colums separated by "|;#" I need to search the file extracting all columns starting with the value of "1" or "2" saving in a separate file just the first 7 columns of each row maching the criteria, with replacement of the saparators in the nearly created... (4 Replies)
Discussion started by: capnino
4 Replies

8. UNIX for Dummies Questions & Answers

Select only certain lines from file and mantain formatting

I want to take the below data, and have it output to file only the STMC#/(IP address) and the "there are X number of updates to install" lines for each machine. I know it's easy, but Im a beginner in BASH stuff, my solution would probably take way too many lines to do something easy.Thanks! ... (5 Replies)
Discussion started by: glev2005
5 Replies

9. AIX

Randomly appearing control characters in text files

Hi, From some time, we have noticed that our ascii files have started corrupting due to the presence of some random control characters (^@, ^M, ^H, ^D). The characters appear randomly on any file after the process that creates the file finishes. If we rerun the process, the files re creates... (0 Replies)
Discussion started by: aakashahuja
0 Replies

10. Shell Programming and Scripting

how to select a value randomly

on my desktop i am using the kde rotating desktop image option. this rotates images randomly every half hour. now, i would like to write an html file which will have an inline frame with some text, maybe system messages, or my friends live journal thati read alot, or unix.com! however, i dont want... (1 Reply)
Discussion started by: norsk hedensk
1 Replies
Login or Register to Ask a Question