Hi guys, Some help with AWK would be appreciated

[GlenA]

Hey, it's not urgent but i would really like to have this working.

We just got our new linux server in the office and we were previously working on NT. With NT i could extract files into a comma dilimited form and then just search the records, whereas with Linux i'm having a few problems, i've asked a few people around and they said that with AWK i should be able to extarct all that i need from a text file.

Well, here's the question: would that be possible to write an AWK application that would display and extarct entries according to the first character of the field? So for every line of input in a file it will find words starting with "N or #". Count the number of letters/digits in each of these words and at the end print out the average number of letters per word beginning with "N or #"?


Thank you

 

[vgersh99]

yes, it would. Use 'match' and associative arrays

vlad
+----------------------------+
| #include<disclaimer.h> |
+----------------------------+

 

[GlenA]

I believe i 've found an easier solution, unfortunately only for a first bit

awk '$0 ~ /[N]/{
for (i = 1; i <= NF; i++) {
if ($i ~ /^[N]/) {
print $i, length($i)
}
}
}'




how do i go about: at the end of output, print out the total number of letters (for words starting with N and the average number of letters per words beginning with &quot;N&quot;?.

Thanks

 

[CaKiwi]

Something like this.

awk '$0 ~ /[N#]/{
for (i = 1; i <= NF; i++) {
if ($i ~ /^[N#]/) {
print $i, length($i)
len += length($i)
k++
}
}
}
END{
if (k) print &quot;ave = &quot; len/k
}'


CaKiwi

&quot;I love mankind, it's people I can't stand&quot; - Linus Van Pelt