what does this awk script do?

sharathkv25 · 03-08-2007, 01:14 AM

Hi,

I would like to know what this awk script does?

filename: bt.as

Code:

BEGIN   {
                "echo $ORA_DB" | getline oracle_home;
                "echo $ORA_APP" | getline oracle_home;
                "echo $ORA_BASE" | getline oracle_base;
        }
/\$ORA_DB/ {gsub("\$ORA_DB",oracle_home);}
/\$ORA_BASE/ {gsub("\$ORA_BASE",oracle_base);}
{print $0 }

this is invoked as:

Code:

awk -f bt.as $PRODUCT_DIR/osmf-http._env > $PRODUCT_DIR/osmf-http.env.bt

druuna · 03-08-2007, 07:16 AM

Hi,

The script will probably not do that much, both gsub statements are incorrect.
It should be: gsub(/string/,replacement), not gsub("string",replacement).

What the maker probably tried to do is:

Get the content of environment variables called ORA_DB, ORA_APP, ORA_BASE and stores these in oracle_home and oracle_base (Strange: both ORA_DB and ORA_APP are stored in oracle_home, the second overriding the first, which is _probably_ incorrect/unwanted).

After this is done, all lines in $PRODUCT_DIR/osmf-http._env are checked for the existence of \$ORA_DB and/or \$ORA_BASE (the literal name, not the content) and substitute these with the content found in the environment variable.

I.e:
If ORA_DB holds /path/to/db, oracle_home is set to /path/to/db
If $PRODUCT_DIR/osmf-http._env has the following line: $ORA_DB/http/specific/db it will be changed to:
/path/to/db/http/specific/db

But like I stated before: The script isn't written correctly and will probably not work.

Hope this clears things up a bit.

cfaj · 03-08-2007, 02:32 PM

Quote:

Originally Posted by druuna

Hi,

The script will probably not do that much, both gsub statements are incorrect.
It should be: gsub(/string/,replacement), not gsub("string",replacement).

More accurately, it is gsub(/regexp/,replacement), and there is nothing wrong with gsub("string",replacement):

Code:

$ echo qwertyuiopqwertyuiop | awk '{gsub( "rty","ZXC" ); print }'
qweZXCuiopqweZXCuiop

druuna · 03-08-2007, 03:10 PM

Hi,

Gnu awk is forgiving and using "..." instead of /.../ will probably have the same results. But...

Other awk flavors (awk/mawk/nawk to name just 3) are stricter/different and only work when using the correct syntax (as found in many awk documents/books).

It's always good practise to use syntactically correct commands, these will work in all flavors. All other extra 'stuff' is great, but will probably bind you to one specific flavor.

gsub(/regexp/,replacement) => works on gnu awk/awk/nawk (don't have mawk around, but it will probably work) and is in general syntactically correct.

gsub("string",replacement) => will work with gnu awk and (BSD)awk, will not work with nawk and (SUN) awk.