> > "Can bash please implement multidimensional arrays as I think they're
> > nifty and would like to use them."
> It seems that Chet has never been interested, and no one else has
> stepped up to contribute.
A large reason for this is because bash is a shell, not a programming
language. Another reason is that there are already ways to work
around the "issue".
The standard way to implement something that works like a multidimensional
array is to use an associative array, and construct a key string out of
your multiple indices, using some delimiter character that can't appear
in any of the indices.
For example, if your indices are integers, you could use a comma
as the delimiter. Then:
declare -A grid=( [0,0]=foo [0,1]=bar [0,2]=baz )
x=0 y=2
echo "${grid[$x,$y]}"
This prints "baz" as expected.
Another way which *only* works when your indices are small non-negative
integers is to use a sparse indexed array, and construct a key index
of the form i + (A)j + (A^2)k + ... where A is some suitably large
constant chosen for your particular problem. For example, to store
up to a 100x100 grid, we can choose A = 100, and let the individual
indices run from 0 to 99.
unset grid
grid=( [0+100*0]=foo [0+100*1]=bar [0+100*2]=baz )
x=0 y=2
echo "${grid[x+100*y]}"
This has the advantage of working in bash versions 2 and 3 which lack
associative arrays, and the disadvantages of requiring numerical indices,
and a strict up-front limit on the dimensions of your matrix.
If you can't abide using "hacks" to implement your own multidimensional
arrays, then bash isn't the right language for your project. Choose a
different one.
