Shell Programming and Scripting

I want to append values to same string variable inside a recursive function that I have .. I do not want to write to any file but use a variable..
Can anyone please help with it? Thanks in advance.

Code:

 
$ echo ${a}
test

$ a=${a}TEST

$ echo $a
testTEST

hi Itkamaraj,

Thanks for reply but this wont serve my purpose .. As i need to use that inside a recursive loop.
Let me provide example of what i need to do..
I have data in form of index but in words not numbers. I need to write that as shown
I need to go down one level till i find _ at end of string , that is why using recusive function. I will keep appending in the string variable with each call to recursive function and return to previous level once there is no more _'s.

Sample DATA::

Code:

level1   level2     level3     level4
Maths_
         geometry_
                      circles
                      triangles_
                                  equilateral-triangles
                      squares
         number theory_
                      primes
English_
         Nouns

OUTPUT
I need to write it in form:

Code:

maths
maths/geometry
maths/geometry/cirlces
maths/geometry/triangles
maths/geometry/triangles/equilateral-triangles
maths/geometry/squares
maths/numbertheory
maths/numbertheory/primes
:
:
English
English/nouns

------------------------------------------------------------

I have the function body i need help with defining the variable string .
Thanks in advance.

Moderator's Comments:

Please use code tags!

Can anyone please brainstorm and help me ?

I gave up the idea of using string variable as it keeps growing larger with each recursive function the control goes in rather than maintaining a value when control comes back at a given level .i.e.

Code:

stringvar="Maths/geometry/circles/triangles/equilateral-triangles"

I am trying to implement it by using array variable but Cant seem to make head way as of now ?

Code:

# awk '/^[Aa-zZ]*_$/{delete d;k=0};NR>1{gsub(/ */,"",$0);
if(match($0,"[Aa-zZ]*[Aa-zZ]*_$",a))
{d[x]=d[x]"/"a[0];sub(/_/,"",d[x]);sub(/^\//,"",d[x]);gsub(/ */,"",d[x]);k++;print d[x];w=0;}
else{w++;for(i=1;i<w;i++){sub(/\/[Aa-zZ]*$/,"",d[x])};d[x-1]=$0;sub(/ */,"",d[x-1]);
print d[x]"/"d[x-1]}w}{for(i=1;i<w;i++){sub(/\/[Aa-zZ]*$/,"",d[x])}}' infile
Maths
Maths/geometry
Maths/geometry/circles
Maths/geometry/triangles
Maths/geometry/triangles/equilateral-triangles
Maths/geometry/squares
Maths/numbertheory
Maths/numbertheory/primes
English
English/Nouns

Dear ygemici,
Can you explain what awk is doing.

Thanks in Advance.
Regards,
Ajay

firstly i m not be sure that's the way is best solution but i just tried something.
well,let me try to explain as much as i ..
i removed "k" variable because of it's unnecessary..maybe more revisions can be done..

Code:

"/^[Aa-zZ]*_$/"

-> [[ if pattern has only like "string_" then delete array named "d".
[[part of `delete array` actually will be used later for new starts (for example Maths_ or "English_" string]]

Code:

"NR>1"

-> [[ when the line number is greater than 1,then process on the records ]]

Code:

"gsub(/ */,"",$0);"

-> [[ if the record ($0) has spaces then remove its. ]]

Code:

"if(match($0,"[Aa-zZ]*[Aa-zZ]*_$",a))"

-> [[ if pattern matchs "string_" ]]
then
[[ a[0]=string_ and create array named "d" and d[x] elemnt = d[x] and a[0] (a[0] equals "string_" like "geometry_") ]]

Code:

sub(/_/,"",d[x]); -> [[ remove "_" ]]
sub(/^\//,"",d[x]); -> [[ remove "/" from beginning of the record ]]
gsub(/ */,"",d[x]); -> [[ remove all spaces if has ]]

Code:

print d[x];w=0;}

-> [[ print d[x] element and w is zero ( just think as a write/print control) ]]

Code:

else
{w++;

-> increase `w` (w is actually both print_control and index for added records (from d[x]=d[x]"/"a[0] )

Code:

for(i=1;i<w;i++)

-> [[ if we dont have "string_" (like "circles" line) my goal must be remove last of the array element (d[x]) (`w` times)
here is what `w` for index for "with "_" or non"_" strings ]]

Code:

{sub(/\/[Aa-zZ]*$/,"",d[x])};

-> [[ remove end of the array element
[[for example my goal must be `Maths/geometry/squares` to `Maths/numbertheory`
so must remove 2 times and add ("numbertheory") to end of array element []d[x][]
[[(actually array is not mandatory to use,but i like arrays )]]

Code:

d[x-1]=$0;

-> [[ and new record is d[x-1] [another array element]]]

Code:

sub(/ */,"",d[x-1])

-> [[remove spaces if has ]]

Code:

print d[x]"/"d[x-1]}w}

-> [[print both d[x] and d[x-1] [[ end of if loop ]]

Code:

{for(i=1;i<w;i++){sub(/\/[Aa-zZ]*$/,"",d[x])}}

-> [[ remove last of the array element (d[x]) (`w` times) from the d[x]
so we get a new array with main d[x] and re-start from it for same operations and both states which for new or else (else portion of if-else) ]]

regards
ygemici