Shell Programming and Scripting

Not sure if the title of this thread makes sense, but hopefully my explanation will.

I'm using awk to print some stats from an apache accesslog. I would like to specify the regexp condition where only the two root pages of "index.html" and "/" are counted in my results. What I can't figure out is how, using regexp or otherwise, I can specify that I only want the forwardslash entries counted ONLY if the forwardslash is at BOTH the start and end of the line....in otherwords, it's the ONLY character on that line. Here's what I have that I thought would work but isn't:

Code:

awk '$7 ~ /^\/index.html|^$\// {print $7}' access_log |sort |uniq -c

I suppose I can't expect the machine to interpret my placing ^ and $ next to one another to mean "this string is at the beginning AND end of the line".

Any ideas? Thanks all.

It'd be better if you post some sample of your data, in the meantime to retrieve / count the lines that start and end with a forward slash, you can try this,

Code:

awk '/^\/index.html/ && /\/$/ { print ... }' ...

Rubin,

Thanks for the direction. Here are a few lines (field 7 only) of the accesslog I'm working with:

/ (count lines like this one)
/index.html (count lines like this one)
/
/
/portals/index.html (don't want lines like this one counted)
/family_medicine/index.html
/
/
/portals/pda/ (don't want lines like this one counted)
/portals/surgery/
/portals/pediatrics/


Where each line that contains "/" and "/index.html" are the only lines I'm interested in. Note: I'm only interested in the root level index.html not the cases like "/portals/index.html".

Using your code, I'm not finding any matches. When trying just one of the conditions, such as:

Code:

awk '$7 ~ /^\/index.html/ {print $7}' access_log-20080213 |sort |uniq -c

it works fine-- I'm showing a count of 5 hits on "/index.html". However, when I try your example of combining the two conditions, that's where I'm not getting any results at all. If I try ONLY your added condition (awk '$7 ~ /\/$/ {print $7}' ....), I get results including lines such as "/portals/pediatrics/" which I don't want.

The AND/OR logic is a touch confusing for me here because I definitely want the count incremented for BOTH conditions, but by using AND it's not going to take it that I mean that field 7 must contain BOTH "/" and "/index.html" per line will it? Obviously OR may not apply here because I don't want it choosing between the two conditions-- I want them both counted.

I hope this clarified and didn't complicate! thanks again.

Derek

One way ...

Code:

awk '$7 == "/index.html"  ||  $7 == "/" {print $7}' file