awk append fileA.txt to growing file B.txt

Shell Programming and Scripting
1

This is appending a column.
My question is fairly simple. I have a program generating data in a form like so:

1  20
2  22
3  23
4  12
5  43

For ever iteration I'm generating this data. I have the basic idea with

cut -f 2 fileA.txt | paste -d  >> FileB.txt ????

I want FileB.txt to grow, and my program automatically generates a new filea.txt after every iteration. Thanks for your help guys!

2

Hi theawknewbie,

I don't understand what you try to achieve. What content will be in fileB?

Regards,
Birei

3

I want FileB to be growing after each iteration.
After 1 iteration, here is FileB:

1  20
2  22
3  23
4  12
5  43

Second iteration, i cut column 2 from the newly made fileA into FileB:

1  20  21
2  22  23
3  23  24
4  12  25
5  43  21

And after the third iteration, fileA is again regenerated, i cut column two from it, append to FileB:

1  20  21  14
 2  22  23  15
 3  23  24  12
 4  12  25  23
 5  43  21  38

etc. etc. My problem is that If i paste the columns, i have to give the file a new name. But I want FileB to maintain the same name, essentially to keep appending and appending to FileB.

4

There will be better ways. With this method, create first fileB. And 'awk' command uses an auxiliar file (fileC in my example). Test it:

$ ls -1
fileA
$ cat fileA 
1  25
2  27
3  28
4  17
5  48
$ touch fileB
$ ls -1
fileA
fileB
$ awk 'NR == FNR { $1 = ""; sub( /[[:space:]]+/, "", $0 ); fileA[FNR] = $0; next } { print $0 "  " fileA[FNR]; fileB_exists = 1 } END { if ( ! fileB_exists ) { for ( i = 1; i <= length(fileA); i++ ) { print fileA } } } ' fileA fileB >fileC; mv fileC fileB
$ cat fileB
25
27
28
17
48
$ awk 'NR == FNR { $1 = ""; sub( /[[:space:]]+/, "", $0 ); fileA[FNR] = $0; next } { print $0 "  " fileA[FNR]; fileB_exists = 1 } END { if ( ! fileB_exists ) { for ( i = 1; i <= length(fileA); i++ ) { print fileA } } } ' fileA fileB >fileC; mv fileC fileB
$ cat fileB
25  25
27  27
28  28
17  17
48  48

Regards,
Birei

5

Birei,

that is sort of the way I was doing it. But the files I'm dealing with become huge, and it's becoming very slow.

I wonder if there is a way to do it in sed?