[SOLVED] Bash - string1 replaces part of string2 on concat

yzfr1 · 04-18-2012, 11:44 AM

Hello there,

This has confused the hell out of me for a while now, so here I am.

I have three arrays. Two with numbers and one with text. I'm trying to neatly print these three arrays out next to eachother, but here's where I get my problem.

The arrays look like this:

Code:

tsizes		fsizes		files
-----		-----		-----
1462417408	735940608	File 1
578531636	367489024	File 2
578517728	367593472	File 3
578452724	366977024	File 4
578439568	367568896	File 5
578448102	365772800	File 6
578424864	365948928	File 7
578448934	369041408	File 8
578452528	367218688	File 9
578440336	367177728	File 10
578421354	366243840	File 11
578425152	367912960	File 12
578445224	367144960	File 13
578432682	367065088	File 14
578439098	367116288	File 15
578430258	367112192	File 16
577640484	367028224	File 17
578073694	367198208	File 18
578432472	367226880	File 19
578448496	367321088	File 20
578438178	367347712	File 21

Now I'm trying to print it out in pretty much the same way as you see it above, separated with " | ". Printing the out one by one is no problem, but when I want them next to each other the problems start.

Here was my first attempt:

Code:

for (( i=0; i< ${#tsizes[@]}; i++ ))
do
	echo "${tsizes[$i]} | ${fsizes[$i]}$ | {ffiles[$i]}"
done

And here is the output from that:

Code:

 | 735940608 | File 1
 | 367489024 | File 2
 | 367593472 | File 3
 | 366977024 | File 4
 | 367568896 | File 5
 | 365772800 | File 6
 | 365948928 | File 7
 | 369041408 | File 8
 | 367218688 | File 9
 | 367177728 | File 10
 | 366243840 | File 11
 | 367912960 | File 12
 | 367144960 | File 13
 | 367065088 | File 14
 | 367116288 | File 15
 | 367112192 | File 16
 | 367028224 | File 17
 | 367198208 | File 18
 | 367226880 | File 19
 | 367321088 | File 20
 | 367347712 | File 21

Ie, the first array isn't printed.
However, having the first array printed on its own in the loop does work. And when I try adding some random characters to print alone with the first array things get really weird:

Code:

for (( i=0; i< ${#tsizes[@]}; i++ ))
do
	echo  "${tsizes[$i]}test"
done

Output:

Code:

test417408
test31636
test17728
test52724
test39568
test48102
test24864
test48934
test52528
test40336
test21354
test25152
test45224
test32682
test39098
test30258
test40484
test73694
test32472
test48496
test38178

The word test is replacing the first four numbers!

Does anyone have an idea on why it would do that and how I can fix it?

Cheers!

grail · 04-18-2012, 12:09 PM

I would hazard a guess that the file where the 'arrays' are being read from was created in Windows and hence the line endings are incorrect.
Try running dos2nix over it or if not sure before you start try:

Code:

cat -A file

If they do not end in a $ sign then linux will not understand them (that is output of above command will show $ signs at the end if correct).

yzfr1 · 04-18-2012, 12:21 PM

Damn, you are so right!

One of the arrays is part of a windows .txt file from a windows machine, read by

Code:

cat "$1" | cut -s -f3

I didn't know it would cause such problems just reading one column from the file, but there you go! Learn something new every day.

Thanks mate!

Nominal Animal · 04-18-2012, 03:34 PM

You could use

Code:

awk 'BEGIN { ORS="\n"; RS="[\t\v\f ]*(\r\n|\n\r|\r|\n)[\t\v\f ]*"; FS="[\t\v\f ]+" }
    { print $3 }' "$1"

instead of cat and cut; it accepts any newline convention, ignores leading and trailing whitespace, and outputs using Unix newlines only. You can easily add filters, too. For example, to do the same, but skip lines that begin with a # or a ;

Code:

awk 'BEGIN { ORS="\n"; RS="[\t\v\f ]*(\r\n|\n\r|\r|\n)[\t\v\f ]*"; FS="[\t\v\f ]+" }
    /^[#;]/ { next }
    { print $3 }' "$1"

yzfr1 · 04-18-2012, 04:06 PM

Really need to learn some of that awk magic some day.

Thanks for the alternative mate, appreciate it.

David the H. · 04-19-2012, 11:53 AM

To add a couple of suggestions for your loop:

1) Assuming all the arrays have an equal number of entries (or you know which one will have the largest number), you can use "${!array[@]}" to output a list of all existing index numbers. This is particularly handy when you have sparse arrays (with missing indexes).

2) The '[]' index boxes in regular arrays (but not associative arrays) operate in an arithmetic environment, so you generally don't need to use '$' on variables. They'll expand automatically.

Code:

for i in "${!tsizes[@]}" ; do
	echo "${tsizes[i]} | ${fsizes[i]} | ${ffiles[i]}"
done

How can I use array variables?
http://mywiki.wooledge.org/BashFAQ/005