matching substrings to multiple part of a line

verse123 · 10-28-2011, 03:49 PM

Hi guys,

I want to find to print out lines of a chart only if a substring matched in multiple parts of that line. So for ex:

col1 col2 col3 col4
this-dog 0-dog that-dog that-dog
this 4 that-dog that-dog

In this example I would only print Row 2 because the substring dog is present in all instances. I have tried something like this, but it has not worked:

Code:

 awk '{if ($1 == /.dog/ && $2 == /.dog/ && $3 == /.dog/ && $4 == /.dog/){print}} file > output

Why won't this work (I don't get an error, I just get a file size of 0) and how can I fix it?

Thanks!

druuna · 10-28-2011, 04:06 PM

Hi,

Your syntax is not correct, have a look at this:

Code:

 awk '{ if ( $1 ~ "dog" && $2 ~ "dog" && $3 ~ "dog" && $4 ~ "dog") { print } }' file > output

Hope this helps

GnuAWK - Comparison Operators

trbennett48 · 10-28-2011, 04:08 PM

Try this... the number of fields it processes is based
NF (number of fields) which is set automatically by awk:

Quote:

awk '{ \
nbr_misses = 0; \
for(idx=1;idx<NF;idx++) \
{ \
loc = index($idx, "dog"); \
if ( loc == 0 ){ nbr_misses++; }
} \
if ( nbr_misses == 0 ) \
{ \
print; \
} \
}'

verse123 · 10-29-2011, 05:43 PM

Thank you for your help and suggestions - they worked!

If I wanted to take this further and only have awk print if a substring matches exactly 3 times in a line, how would I do this?

with the code that Druuna suggested (the simpler one) it will print output even if "dog" is only found in 1/3 lines (or 2/3 lines, but not exactly 3).

grail · 10-30-2011, 01:49 AM

You don't seem to mention what if it is more than 3, but here is an alternative:

Code:

awk 'match($0,/(dog).*(dog).*(dog)/,f) && f[3]' file