Need to get the file name and put this in the file itself

[smehta]

I need to do the following. I have the following file called
emp-name-irving-kentucky.grammar

The original contents are:
([
(elbert freeman){<vid bkaa9v2>}
(marque worrell){<vid b3ygmys>}
(ronney riddell){<vid b3zbgjq>}
]
?please~0.001)

I need the file to be transformed to the followinG;
((irving kentucky)[
(elbert freeman){<vid bkaa9v2>}
(marque worrell){<vid b3ygmys>}
(ronney riddell){<vid b3zbgjq>}
]
?please~0.001)

So basicallly I am taking the city and state out of filename
emp-name-irving-kentucky.grammar and putting it as
(irving kentucky) on the top of the same file....

How can this be done?

Also once I am done with file I need to do this for all file under one directory and concatenate all of these
city state files to produce one file.

 

[vgersh99]

For one file:

nawk -f townState.awk emp-name-irving-kentucky.grammar


#------------------------ townState.awk ----------------

BEGIN {
split(FILENAME, townState, &quot;-&quot;
town=townState[3];
state=substr(townState[4], 1, length(townState[4]) - index(townState[4],&quot;.&quot+1
);

# printf(&quot;town->[%s] state->[%s]\n&quot;, town, state);
}

/^\(\[/ { printf(&quot;((%s %s)[\n&quot;, town, state); next }

{ print }

#--------------------------------------------------

For multiple file, write a shell script to loop through all the desired files. SOmething like that:

#!/bin/ksh

for iter in `ls *grammar`
do
nawk -f townState.awk ${iter} > ${iter}NEW
done;


vlad

 

[bigoldbulldog]

Great script Vlad!

Here's a fix and an addition. The fix is for getting the state. It happened to work for Kentucky but not for Ohio. The second is a way of applying this to multiple files in pure awk and saving to new files directly. This could go straight to stdout by taking out the &quot; > outfile&quot; and, optionally, removing the setOutput call and function.

/^\(\[/ {
setOutput()
setTownState()
printf(&quot;((%s %s)[\n&quot;, town, state) > outfile
next
}
{ print > outfile }
function setOutput(){
if( NR > 1 ) close(output)
outfile=substr(FILENAME,1,index(FILENAME,&quot;.&quot) &quot;new&quot;
}
function setTownState(){
split(FILENAME, townState, &quot;-&quot;
town=townState[3];
state=substr(townState[4], 1, index(townState[4],&quot;.&quot-1);
}
Cheers,
ND

bigoldbulldog@hotmail.com

 

[vgersh99]

ND,

thanks for the fix. Here are 2 fixes to your fixes
I think they are appropriate:

1. Append output to a file
2. setOuput ONLY if at the first record if a current file
3. Derive town/state ONLY if at the first record in a current file.

NOTE: I guess the whole thing can be rewritten with the &quot;first-record-in-the-current-file&quot; logic as that's the boundary condition for all other changes.

ND, thanks again for the fixes. BTW, is the &quot;sed&quot; implementation version coming up? [just kiddin']

vlad


/^\(\[/ {
setOutput()
setTownState()
printf(&quot;((%s %s)[\n&quot;, town, state) >> outfile
next
}
{ print >> outfile }

function setOutput(){
if( FNR == 1 ) {
close(outfile)
outfile=substr(FILENAME,1,index(FILENAME,&quot;.&quot) &quot;new&quot;
}
}
function setTownState(){
if (FNR == 1)
split(FILENAME, townState, &quot;-&quot;

town=townState[3];
state=substr(townState[4], 1, index(townState[4],&quot;.&quot-1);
}

 

[bigoldbulldog]

That's great! I always forget about NFR.

One more fix: functions setOutput and setTownState should always be called at line 1 of each file in case /^\(\[/ doesn't occur on line 1 of each. To get:

FNR == 1 {setOutput(); setTownState()}
/^\(\[/ {
setOutput()
setTownState()
printf(&quot;((%s %s)[\n&quot;, town, state) >> outfile
next
}
{ print >> outfile }

function setOutput(){
close(outfile)
outfile=substr(FILENAME,1,index(FILENAME,&quot;.&quot) &quot;new&quot;
}
function setTownState(){
split(FILENAME, townState, &quot;-&quot;
town=townState[3];
state=substr(townState[4], 1, index(townState[4],&quot;.&quot-1);
}

p.s. sed would be easy, but needs help from the shell to get the input filename(s). Maybe when I have time I'll see if I'm right about the 'easy', if someone doesn't beat me to it. Cheers,
ND

bigoldbulldog@hotmail.com

 

[smehta]

These are good but they replace all the occurences of the character ([ with file name and I just need this to be done only once at the begining of the file. Thanks alot anyways. Could anyone suggest the modification