Replacing individual characters with a pattern.

Shell Programming and Scripting
1

Given a paragraph of text which individual words have been prepended with an equals sign, I want to replace the individual characters in each of the marked words with blank spaces (two underscores and a space) for students to fill in -- a cloze passage that lets the students know how many letters are in each of the words they are required to fill.

As an example, something like:

Truth be told, this is the =closest
 I've =ever =gotten to a =college =graduation.

should give:

Truth be told, this is the __ __ __ __ __ __ __  
I've __ __ __ __     __ __ __ __ __ __  to a 
__ __ __ __ __ __ __ __ __ __.

another variation I would like to be able to get is :

Truth be told, this is the c __ __ __ __ __ __  I've e __ __ __  
g__ __ __ __ __  to a g __ __ __ __ __ __ __ __ __.

I'm not quite sure how to go about this but I suspect awk will be the tool to use.

2

What have you done so far?
The second version would be easier...

How would you know from when or what is to be substituted?

3

Try this...

#Variation 1
awk '{ for(i=1;i<=NF;i++) {
                if($i~"=") {
                        l=length($i)-1
                        if($i~"[\\.,]$"){
                          l=l-1; e=substr($i,l+2,l) }
                        for(j=1;j<=l;j++) p=p"_ ";
                        $i=p;printf $i e OFS;p="";e=""
                } else { printf $i OFS }
        }printf "\n" } ' input_file

#Variation 2
awk '{ for(i=1;i<=NF;i++) {
                if($i~"=") {
                        l=length($i)-1
                        if($i~"[\\.,]$"){
                          l=l-1; e=substr($i,l+2,l) }
                        s=substr($i,2,1);
                        for(j=2;j<=l;j++) p=p"_ ";
                        $i=p;printf s $i e OFS;p="";e=""
                } else { printf $i OFS }
        }printf "\n" } ' input_file

--ahamed

1 Like
4

Another variation:

#  cat x.txt
Truth be told, this is the =closest
 I've =ever =gotten to a =college =graduation.
# awk '{for (i=1;i<=NF;i++) { if ($i ~ /^=/) { $i=substr($i,2); gsub ("[[:alpha:]]","__ ", $i) }} print}' x.txt
Truth be told, this is the __ __ __ __ __ __ __
I've __ __ __ __  __ __ __ __ __ __  to a __ __ __ __ __ __ __  __ __ __ __ __ __ __ __ __ __ .
2 Likes
5

ahhhh!!!

I was trying everything for this gsub ("[[:alpha:]]","__ ", $i) .
Awesome code CarloM :slight_smile: And thanks for the info!
My code looks like junk! :smiley:

--ahamed

6

this is fun!

perl -pe 's/(?:=)(\w+)\w/\1__ /g ; 1 while s/(\w)(\w)((__ )+)/\1__ \3/' < input-text-file 

Truth be told, this is the c__ __ __ __ __ __ 
 I've e__ __ __  g__ __ __ __ __  to a c__ __ __ __ __ __  g__ __ __ __ __ __ __ __ __ .

[/code]

1 Like
7

Work brilliantly. Thank you.

---------- Post updated at 01:17 AM ---------- Previous update was at 01:11 AM ----------

Works brilliantly, too. Just one question. I use the bash shell and can usually repeat the previous command by using the up arrow. This command doesn't show up. When I press the up arrow, the command before it appears. What gives?

---------- Post updated at 01:23 AM ----------

[COLOR="\#738fbf"][SIZE=1]---------- Post updated at 01:26 AM ---------- Previous update was at 01:23 AM ----------

Works as advertised. Thanks. Pity I am illiterate in perl.

8

Dunno. Line-wrapping?

Also, modified for your second variation:

# awk '{for (i=1;i<=NF;i++) { if ($i ~ /^=/) { x=substr($i,2,1); $i=substr($i,3); gsub ("[[:alpha:]]","__ ", $i); $i=x" "$i }} print}' x.txt
Truth be told, this is the c __ __ __ __ __ __
I've e __ __ __  g __ __ __ __ __  to a c __ __ __ __ __ __  g __ __ __ __ __ __ __ __ __ .
9

Thanks for the second variation. This one shows up in the shell history. Go figure. :rolleyes: