UNIX for Dummies Questions & Answers

I almost reach my objective (Youhouuu !!!!)
But I really don't understand why it doesn't work until the end...

For clarity's sake I am taking a very simple example. The operations I am doing in the script (gsub and print) really don't have any importance !!! I just matter about redirecting the file that I produced inside the script.


input file called "original.txt":

Code:

book # 001|line 1|America
book # 002|line 56|Asia
book # 003|line 38|Europe

I remove the expression "line " in $2 and would like to redirect the output file called "for_later.txt" because I will need to have a trace of the data I produced at this stage and I will use them later for something else (something even totally different from programming).

But for now, I need to keep processing the original input file "original.txt" by adding another field ("Price"), like if I've never redirected anything.

Here what I tried:

Code:

BEGIN{FS=OFS="|"}

{ 
    sub("line ","",$2)
    print $0 > "for_later.txt"   # at this stage of the script I redirect the output that I will need later
}

{
    while((getline < "for_later.txt") > 0)   # here I redirect the produced file to keep processing it
    print $0 FS "Price:   $"
}

Problem:
The file "for_later.txt" looks perfectly fine:

Code:

book # 001|1|America
book # 002|56|Asia
book # 003|38|Europe

However my final output produced when awk finish executing the entire script only display the first line (the content of this line is correct though):

Code:

book # 001|1|America|Price:   $

Question: Why the 2 other lines are missing ???
the final output I would like for now:

Code:

book # 001|1|America|Price:   $
book # 002|56|Asia|Price:   $
book # 003|38|Europe|Price:   $

* I also tried every kind of getline:
getline
getline var
getline < file
getline var < file
command | getline (with the shell command "cat for_later.txt")
command | getline
command |& getline
command |& getline var

* I tried to close the file before calling getline.

* I tried "if" instead of "while", and even with no statements.

It always return the first line only !

Please, if someone could explain what is happening it would be great.
Thanks !

Your description of the problem does not mention anything about for_later.tab but you use it in your script. What does it look like?

Whoops sorry, in the code I meant "for_later.txt" (I corrected the code). There is no .tab file

Actually when I tried your code in gawk, I did not get any terminal output. That may be because gawk buffers its output. So, the file for_later.txt actually has records only when the program finishes (or when the buffer is flushed) and not in between.

But, I could simulate the behaviour mentioned by you by putting in a system call with a null string (system("")) after print $0 > "for_later.txt". This flushes the output buffers to their destinations.

When I tried your code with the awk which comes with AIX 6.1, I got what you want to see:

Code:

book # 001|1|America|Price:   $
book # 002|56|Asia|Price:   $
book # 003|38|Europe|Price:   $

I am using gawk as well and I get only the first line on the terminal output.

Code:

2) the while loop tries to read a line at a time from for_later.txt until an error. This reads only the first line (output by 1) and prints it.

So, is there any way to "force" or bypass the error exit status of getline and set it to 1 in order to keep reading the remaining line of "for_later.txt"?

Why don't you modify your script to this?:

Code:

BEGIN{FS=OFS="|"}

{ 
    t=$0
    sub("line ","",$2)
    print $0 > "for_later.txt"   # at this stage of the script I redirect the output that I will need later
    print $0,"Price:   $"  
}

As I mentioned in my first post, this is a simple example to show the problem of redirecting the entire file into the script.

The purpose of this thread is to know how I could do that with getline or other strategies.

But thanks anyway !

So I presume it is not possible.