Hi.

In a language such as C or something you can define a variable to be of type double, say, and then go on and perform numerical operations involving this variable.

result=var*3.14 for example, where result would be defined to be of type double also.

Are such operations possible within shell scripts? I've defined a variable $norm to represent the number of seconds in a day in a bash script:

norm=24*60*60

But then just to check everything was as it should be I echoed the variable norm to the screen ( echo $norm ) and it gave me: > 24*60*60

I have a bit of simple multiplying and dividing to do involving variables...is this possible within shell scripts?

Thanks

Yes, you can use:
(( norm = 24*60*60 ))
or
let "norm=24*60*60"

Thanks for that.

But do you know if everything is treated as type double or what?

eg if a & b were defined as type int and were assigned:

a=5
b=2

then a/b=2 in other languages

Are there type definitions in scripting or is it handled automatically somehow?

Integers and strings only. You do not need a type definition:

darktown:~ # (( norm=5/2 ))
darktown:~ # echo $norm
2
darktown:~ # VAR="some text"
darktown:~ # echo $VAR
some text
darktown:~ # VAR=11/2
darktown:~ # echo $VAR
11/2
darktown:~ # (( VAR = 11/2 ))
darktown:~ # echo $VAR
5
darktown:~ # (( VAR = 11%2 ))
darktown:~ # echo $VAR
1

The % operator is for modulo.
If you need floating point, bash is not what you need.

ok. that's cleared it up. Thanks.

bash doesn't have an internal mechanism to do floating point arithmetic. The $((expression)) syntax only does integer math.

You can always put a command through bc -l using the `backtick execution` syntax (or $(like this) if you prefer - as I do):
You could also switch shell to use zsh, which can do floating point arithmetic internally.

Just to mention, you can use something like this too:
Though this is specific to bash probably.