LinuxQuestions.org - (bash) how to list all variables?

- Programming (https://www.linuxquestions.org/questions/programming-9/)

- - (bash) how to list all variables? (https://www.linuxquestions.org/questions/programming-9/bash-how-to-list-all-variables-305017/)

(bash) how to list all variables?

Hi,

how can I list all variables that are sourced from another file?

Code:

#!/bin/bash

. /tmp/variables

"list_all_varibles" > /root/variables_in_use.txt

/tmp/variables looks like this:

A=1
B=2
C=3
#D=5

I think that no matter how you play it, you're going to have to do some work. AFAIK, there is no easy solution. If your variable file will always look like your example, you can write the bash to look for lines that don't start with # and which contain an equals sign (=) and strip off the equals and anything after it.. that would at least provide you with the variables that have been assigned.

That's about the best of which I can think at the moment.

Anyone else?

ddi you want all the variables sourced from a particular file?
if you wanted all the variables in use then you can use
set

And maybe you could store the output of set, source the file, store another output of set, then
compare the outputs

Quote:

Originally posted by zeropash
ddi you want all the variables sourced from a particular file?
if you wanted all the variables in use then you can use
set

I'm not very familiar with the set command, but it seems that this works:

Code:

#!/bin/bash

set -a

. /tmp/variables

env > /root/variables_in_use.txt

edit:
or this seems to work as well (and I think this is what you were talking about):

Code:

#!/bin/bash

. /tmp/variables

set > /root/variables_in_use.txt

Those list also all normal environment variables, but it's ok.

Thanks!

sdiff can help to filter the "unnecessary" stuff out:

Code:

#!/bin/bash

env > env.1

. /tmp/variables

env > env.2

sdiff -s env.1 env.2 |cut -d">" -f2 > /root/variables_in_use.txt

HTH.

Quote:

Originally posted by dustu76
sdiff can help to filter the "unnecessary" stuff out:

Code:

#!/bin/bash env > env.1 . /tmp/variables env > env.2 sdiff -s env.1 env.2 |cut -d">" -f2 > /root/variables_in_use.txt

HTH.

Env doesn't work in that example, but set does.

sdiff seems to clean most of the "useless" variables away.

Thanks!

set | grep -f variables-in-use
this works, but no blanks allowed in the data file

DATA

Code:

billym.primadtpdev>cat ~/1

A=1

B=2

C=3

#D=5

SCRIPT

Code:



billym.primadtpdev>cat ~/2

. ~/1

echo we get:

set | grep -f ~/1

RESULT

Code:

billym.primadtpdev>ksh ~/2

we get:

A=1

B=2

C=3

Quote:

Originally posted by keefaz
And maybe you could store the output of set, source the file, store another output of set, then
compare the outputs

That's pretty much the way I'd handle this.

Code:

set>/tmp/variables.0

. file-containing-variable-defs

set>/tmp/variables.1

echo "file-containing-variable-defs defined the following:"

cat /tmp/variables.0 /tmp/variables.1 | sort | unit -u

rm /tmp/variables.[01]

I suppose env would work just as well in the above (as some others suggested). If you only want the variable names and not their values, just pipe the result from "uniq" into "cut -d= -f1"

I wrote a shell function a long time ago that did something like (I think) Duudson was looking for. It was to allow someone to display the contents of a config file that was being used in a script. It read the config file a line at a time rather than sourcing it and only displayed the non-comment lines before defining the variable. (I didn't want to deal with temp files at the time even though sourcing the file would have been faster.)