Applying Unix's Head on AWK Through Find Command

Question

I want to output top 10 lines of AWK command in the list of files given by find, using this snippet:

$ find . -name "*.txt" -print -exec awk '$9 != ""'  \| head -n10 {} \;

Note also that I want to print out the file names being processed.

But why I get such error:

awk: cmd. line:2: fatal: cannot open file `|' for reading (No such file or directory)
./myfile.txt

What's the right way to do it?

I tried without backslash before the pipe. Still it gave an error:

find: missing argument to `-exec'
head: cannot open `{}' for reading: No such file or directory
head: cannot open `;' for reading: No such file or directory

Please clarify what you are trying to do. It seems like you are trying to achieve two things at once. — Yuval F, Apr 02 '09 at 07:11
@YuvalF: 1. With AWK filter out lines that contain "" in 9th column; 2. Show only top 10 lines - after the filter. — neversaint, Apr 02 '09 at 07:18

score 4 · Answer 1 · answered Apr 02 '09 at 07:41

When running a command with find's -exec, you don't get all the nice shell things like the pipe operator (|). You can regain them by explicitly running a subshell if you like though, eg:

find . -name '*.txt' -exec /bin/sh -c "echo a text file called {} | head -n 15" \;

ashawley · Accepted Answer · 2009-04-06T16:08:47.087

3

If you want to run an Awk program on every file from find that only prints the first 10 lines each time.

$ find . -name "*.txt" -print -exec awk '$9 != "" && n < 10 {print; n++}' {} \;

edited Apr 06 '09 at 16:08

answered Apr 03 '09 at 14:28

ashawley

4,195
1
27
40

score 2 · Answer 3 · answered Apr 06 '09 at 16:44

2

Based on Ashawley's answer:

find . -name "*.txt" -print -exec awk '$9 != "" {print; if(NR > 9) exit; }' {} \;

It should perform better, as we exit awk after the 10th record.

answered Apr 06 '09 at 16:44

MatthieuP

1,116
5
12

It exits after the tenth record, but not after the tenth printed line which is what the OP wanted and ashawley's answer provides. – Dennis Williamson May 23 '09 at 04:05

Zsolt Botykai · Answer 4 · 2009-04-02T08:12:45.633

1

You can do it this way too:

find . -name '*txt' -print -exec awk 'BEGIN {nl=1 ;print FILENAME} $9 !="" {if (nl<11) { print $0 ; nl = nl + 1 }}' {}  \;

without head.

edited Apr 02 '09 at 08:12

answered Apr 02 '09 at 07:21

Zsolt Botykai

50,406
14
85
110

@Zsolt: I tried. For each file, your snippet keeps going on even after 10th line. – neversaint Apr 02 '09 at 07:25
@Zsolt: thanks. How can I make it to print the file name being (as delimiter) also? – neversaint Apr 02 '09 at 07:52

score 1 · Answer 5 · answered Apr 02 '09 at 07:44

1

Using awk only should work:

find . -name "*.txt" -print -exec awk '{if($9!=""&&n<11){print;n++}}' {} \;

answered Apr 02 '09 at 07:44

mouviciel

66,855
13
106
140

@mouviciel: thanks. How can I make it to print the file name being processed also? – neversaint Apr 02 '09 at 07:50

Applying Unix's Head on AWK Through Find Command

5 Answers5