Aargh.. can't figure this out: for( i in s), index

rossb

Following is a script file snippet which is intended to replace a keyword with another string, which contains a numeric (line #) index into a key word file of following form:

keyword;replacetext

...

keyword_n;replacetext_n

It parses all *.htm files in a directory, parses and creates equivalent *.htp files. Parsing and replacement works fine.

#!/bin/sh

INEXT="htm"

OUTEXT="htp"

...

ParseKeys()

{

# $1 = KeyWordFile

find . \( ! -name . -prune \) -type f -name "*.$INEXT" | while read FILE

do

awk 'BEGIN{ FS=";"}

FNR==NR{ s[$1]=$2; next }

{

def=0;

for( i in s ) {

system("echo hello alldef: " def " i: " i " >>trash.txt")

if( $0 ~ i ){

system("echo hello def: " def " >>trash.txt")

system("echo hello i: " i " >>trash.txt")

system("echo hello s[i]: " s[i] " >>trash.txt")

gsub(i,s[i])

}

def=def+1;

}

print $0

}' $1 $FILE > temp

# $FILE is form ./*.htm

outfile=`echo "$FILE" | cut -f2 -d "/" | cut -f1 -d "."`

echo "$outfile"

mv -f temp $outfile.$OUTEXT

done

echo hello

}

The problem is that my line counter (def) for the keyfile is out of sync with the keyword (iterator - i) from "for( i in s )"

I get output like:

hello alldef: 0 i: blue_pill

which should be:

hello alldef: 32 i: blue_pill

since blue_pill is on the 32nd line.

does the "for( i in s )" not start iterating at first line, or am I seriously missing something? Is there a better way to do this?

Thanks for looking...

Bill

gomer

you are printing the contents of the variable "def" which you explicitly set to 0. You should be getting 0 for itteration of the loop, I'm guessing.

rossb

Nope...

def=0;
for( i in s ) {
... dostuff
def=def+1;
}

Init def to zero previous to loop, increment each loop iteration

I am seeing that def is a random appearing value mod(max loop iterations=lines) that bears no discernible relationship to the data (i value for file line n) at any particular loop iteration.

def is NOT constant, it just does not appear to bear any relationship to line #

cjcox

Not sure Bill... but I didn't find your description of what you're trying to do to be clear at all.

Do you have a file of replacement text that looks like:

word;newword
word2;newword2
word3;newword3
etc...

I think that's what you're saying.

But on replace, you want the string replaced with the line number preceding of the line from the file of keyword/replace pairs?

Sometimes, you need to look at things from a different perspective... assume the keyword file is called repl.txt, then consider:

grep -n '.' repl.txt | sed 's,^\([0-9]*\):\([^;]*\);\([^;]*\),s/\\\b\2\\\b/\1 \3/g,' >repl.sed

The sed script file now contains substitute commands for your keyword replaces, so you could call it like:

sed -f repl.sed myhtmlfile.html

I'm hoping that's useful.... but maybe I'm way off.

gomer

Bill, you're right. I missed that line. that line is also no good.

Try this:
def=$(( $def + 1 ))

normapomero

"you are printing the contents of the variable "def" which you explicitly set to 0. You should be getting 0 for itteration of the loop"

Would you be more specific or explain on what you really mean.

Custom Software Package Development

normapomero

I really love this site because of the helpful information posted here.