I have three files named One, Two, Three.
One contains:
1
3
2
Two contains:
4
6
5
Three contains:
7
9
8
When I give the following command:
$sort < One < Two < Three
I get the output:
7
8
9
But when I give the following command:
$sort One Two Three
I get the ouput:
1
2
3
4
5
6
7
8
9
Can anyone please shed light on what exaclty is happening here? Why does the input from 1 and 2 not taken into consideration in the first command?
Your command is the same as:
sort 0<1 0<2 0<3
(file descriptor 0 is standard input)
Redirections are processed in the order they appear, from left to right.
sort command itself cannot see any of those files.
bash open file 1,2,3 at file descriptor 0 one by one.
So the right most one override left ones.
At last, sort read from file descriptor 0 which is bind to file 3.
You can't redirect multiple files with bash. To work around this limitation you could use cat:
cat 1 2 3 | sort
On a side note, zsh supports what it calls mutlios:
zsh$ setopt multios
zsh$ sort < 1 < 2 < 3 > 4 > 5
zsh$ tr '\n' ' ' < 4 < 5
1 2 3 4 5 6 7 8 9 1 2 3 4 5 6 7 8 9
Related
So I have a C program that outputs many numbers. I have to check them all. The problem is, each time I run my program, I need to change seeds. In order to do that, I've been doing it manually and was trying to make a shell script to work around this.
I've tried using sed but couldn't manage to do it.
I'm trying to get the output like this:
a=$(./algorithm < input.txt)
b=$(./algorithm2 < input.txt)
c=$(./algorithm3 < input.txt)
The output of each algorithm program is something like this:
12 13 315
1 2 3 4 5 6 7 8 10 2 8 9 1 0 0 2 3 4 5
So the variable a has all this output, and what I need is
variable a to contain this whole string
and variable a1 to contain only the third number, in this case, 315.
Another example:
2 3 712
1 23 15 12 31 23 3 2 5 6 6 1 2 3 5 51 2 3 21
echo $b should give this output:
2 3 712
1 23 15 12 31 23 3 2 5 6 6 1 2 3 5 51 2 3 21
and echo $b1 should give this output:
712
Thanks!
Not exactly what you are asking, but one way to do this would be to store the results of your algorithm in arrays, and then dereference the item of interest. You'd write something like:
a=( $(./algorithm < input.txt) )
b=( $(./algorithm2 < input.txt) )
c=( $(./algorithm3 < input.txt) )
Notice the extra () that encloses the statements. Now, a, b and c are arrays, and you can access the item of interest like ${a[0]} or $a[1].
For your particular case, since you want the 3rd element, that would have index = 2, hence:
a1=${a[2]}
b1=${b[2]}
c1=${c[2]}
Since you are using the Bash shell (see your tags), you can use Bash arrays to easily access the individual fields in your output strings. For example like so:
#!/bin/bash
# Your lines to gather the output:
# a=$(./algorithm < input.txt)
# b=$(./algorithm2 < input.txt)
# c=$(./algorithm3 < input.txt)
# Just to use your example output strings:
a="$(printf "12 13 315 \n 1 2 3 4 5 6 7 8 10 2 8 9 1 0 0 2 3 4 5")"
b="$(printf "2 3 712 \n 1 23 15 12 31 23 3 2 5 6 6 1 2 3 5 51 2 3 21")"
# Put the output in arrays.
a_array=($a)
b_array=($b)
# You can access the array elements individually.
# The array index starts from 0.
# (The names a1 and b1 for the third elements were your choice.)
a1="${a_array[2]}"
b1="${b_array[2]}"
# Print output strings.
# (The newlines in $a and $b are gobbled by echo, since they are not quoted.)
echo "Output a:" $a
echo "Output b:" $b
# Print third elements.
echo "3rd from a: $a1"
echo "3rd from b: $b1"
This script outputs
Output a: 12 13 315 1 2 3 4 5 6 7 8 10 2 8 9 1 0 0 2 3 4 5
Output b: 2 3 712 1 23 15 12 31 23 3 2 5 6 6 1 2 3 5 51 2 3 21
3rd from a: 315
3rd from b: 712
Explanation:
The trick here is that array constants (literals) in Bash have the form
(<space_separated_list_of_elements>)
for example
(1 2 3 4 a b c nearly_any_string 99)
Any variable that gets such an array assigned, automatically becomes an array variable. In the script above, this is what happens in a_array=($a): Bash expands the $a to the <space_separated_list_of_elements> and reads the whole expression again interpreting it as an array constant.
Individual elements in such arrays can be referenced like variables by using expressions of the form
<array_name>[<idx>]
like a variable name. Therein, <array_name>is the name of the array and <idx> is an integer that references the individual element. For arrays that are represented by array constants, the index counts elements continuously starting from zero. Therefore, in the script, ${a_array[2]} expands to the third element in the array a_array. If the array would have less elements, a_array[2] would be considered unset.
You can output all elements in the array a_array, the corresponding index array, and the number of elements in the array respectively by
echo "${a_array[#]}"
echo "${!a_array[#]}"
echo "${#a_array[#]}"
These commands can be used to track down the fate of the newline: Given the script above, it is still in $a, as can be seen by (watch the quotes)
echo "$a"
which yields
12 13 315
1 2 3 4 5 6 7 8 10 2 8 9 1 0 0 2 3 4 5
But the newline did not make it into the array a_array. This is because Bash considers it as part of the whitespace that separates the third and the fourth element in the array assignment. The same applies if there are no extra spaces around the newline, like here:
12 13 315\n1 2 3 4 5 6 7 8 10 2 8 9 1 0 0 2 3 4 5
I actually assume that the output of your C program comes in this form.
This will store the full string in a[0] and the individual fields in a[1-N]:
$ tmp=$(printf '12 13 315\n1 2 3 4 5 6 7 8 10 2 8 9 1 0 0 2 3 4 5\n')
$ a=( $(printf '_ %s\n' "$tmp") )
$ a[0]="$tmp"
$ echo "${a[0]}"
12 13 315
1 2 3 4 5 6 7 8 10 2 8 9 1 0 0 2 3 4 5
$ echo "${a[3]}"
315
Obviously replace $(printf '12 13 315\n1 2 3 4 5 6 7 8 10 2 8 9 1 0 0 2 3 4 5\n') with $(./algorithm < input.txt) in your real code.
I have two files; 1.txt and 2.txt
1.txt has the following content:
a 1 2 3 4 5
b 4 5 6 7 7
c 4 5 6 7 6
d 6 5 4 3 2
and 2.txt;
b
d
I need to extract those lines from 1.txt whose first fields match the first fields of 2.txt;
b 4 5 6 7 7
d 6 5 4 3 2
I thought a simple join command should work for me:
join 1.txt 2.txt
But unfortunately, the command produces just a couple of lines, even though both files are pretty large.
I cannot figure out what's going on.
I have an example like so:
1 2 3 4 5 6 7 8 9 10 2.2
1 3 2 3 2 3 2 3 2 33 1.1
11 values per line, all single spaced.
The occasional random character thrown in, but that's it. I'm trying to find a way to copy the line in which the last value is less than a some user/predetermined value. Something akin to a 'grep if $last <= 2', but I can't think of one nor can I find one.
Thanks for any help!
Simple awk use case:
awk -v val=2 '$NF < val' file
Output:
1 3 2 3 2 3 2 3 2 33 1.1
I am new to shell scripting and I am trying a simple task of getting the length of a sequence of numbers generated using seq.
With the help of a related post here: How to find the array length in unix shell? I was able to do this -
a=(1 2 3 4 5)
echo ${#a[#]} #length of a
5 #length of a = 5 (This is fine !!)
However when I try to do a similar thing using seq ..
b=$(seq 1 1 10)
echo $b
1 2 3 4 5 6 7 8 9 10
echo ${#b[#]}
1 #the length of b is 1, while I expect it to be 10
Why does this happen ? Are the variable types a and b different? is b not an array ?
I am sure I am missing something very trivial here, help is greatly appreciated.
Thanks
Ashwin
You need to store the output in an array to find the length of the array:
$ b=($(seq 1 1 10))
$ echo ${#b[#]}
10
Saying b=$(seq 1 1 10) doesn't produce an array.
Try
echo ${b[0]}
It will be 1 2 3 4 5 6 7 8 9 10 because all your values are stored in first element of array a as a string.
b=($(seq 1 1 10))
will do what you want.
My colleague has given me a file, in which half of the lines are made of 8 columns of info and the other half are made of the 9th column of info. They are always next to each other, e.g.
1 2 3 4 5 6 7 8
1.1
2 3 4 5 6 7 8 9
1.2
...
a b c d e f g h
abcd
I know how to paste every two lines as one and print them out in Python. But I was wondering if it's possible to do that even more conveniently in BASH?
Thanks guys!
You could use sed or awk, as other answers have mentioned. Those answers are all good.
You could also do this easily in pure shell.
$ while read line1; do read line2; echo "$line1 $line2"; done < input.txt
1 2 3 4 5 6 7 8 1.1
2 3 4 5 6 7 8 9 1.2
Note that whitespace is not preserved.
There's another tool available on most unix-like systems called paste:
$ paste - - < input.txt
1 2 3 4 5 6 7 8 1.1
2 3 4 5 6 7 8 9 1.2
In this case, there's a big space in the first line because paste separates columns using tabs, by default, and the trailing space in the first line of input.txt caused the separating tab to be offset to the next column. You can read paste's man page for options to control this.
Another awk
awk '{f=$0;getline;print f,$0}' file
1 2 3 4 5 6 7 8 1.1
2 3 4 5 6 7 8 9 1.2
And just for the fun of it a gnu awk
awk -v RS="[0-9][.][0-9]" '{$1=$1;print $0,RT}' file
1 2 3 4 5 6 7 8 1.1
2 3 4 5 6 7 8 9 1.2
Here is set the Record Separator to the value in line two.
Then the RT will have the actual separator stored.
try:
awk '{printf "%s%s",$0,(NR%2?FS:RS)}' file
or:
awk 'NR%2{printf "%s ",$0;next}7' file
test:
kent$ echo "1 2 3 4 5 6 7 8
1.1
2 3 4 5 6 7 8 9
1.2"|awk '{printf "%s%s",$0,(NR%2?FS:RS)}'
1 2 3 4 5 6 7 8 1.1
2 3 4 5 6 7 8 9 1.2
kent$ echo "1 2 3 4 5 6 7 8
1.1
2 3 4 5 6 7 8 9
1.2"|awk 'NR%2{printf "%s ",$0;next}7'
1 2 3 4 5 6 7 8 1.1
2 3 4 5 6 7 8 9 1.2
You can sed:
sed 'N;s/\n/ /' file
or awk:
awk 'NF==1{print $0}{printf "%s ",$0}' file