3284 Rate this article:
5.0

Using a HASH with FOREACH

Barrett Sather

The FOREACH loop is a handy tool in IDL, allowing a user to iterate over each element in an array without having to determine the size. To start the loop, the syntax is:

foreach element, variable, index do

The first argument "element" does not have to be defined beforehand, and is the variable that holds the element of the loop for the current iteration. The second argument is the IDL variable that contains the array, list or, hash you want to loop over. The third argument "index" is optional, and contains the index for the current loop. When using a list or array, this third argument is a good way to keep track of the iteration you are currently on.

With a HASH however, the third optional argument will be the key for the current key/value pair, and the first argument, "element" will be the value. This is useful for things like titling your graphics, since the key can be a string. This example below illustrates how to do this.

x = findgen(21)/10 - 1.0

i = 1

data = orderedhash()

data['linear'] = x

data['squared'] = x^2

data['cubed'] = x^3

p = list()

foreach this_data, data, type do begin

  p.add, plot(x, this_data, LAYOUT=[3,1,i], TITLE=type, THICK=3, /current)

  i++

endforeach

(p[0]).COLOR= 'red'

(p[1]).COLOR= 'green'

(p[2]).COLOR= 'blue'

The resulting graphic looks like this:

There a couple things to note in this example:

First, since the third optional argument is a key for the HASH, if you need numeric index for the loop, it has to be taken care of manually by incrementing a value (I used "i") in the loop.

Second, the use of an ORDEREDHASH is important. Without this, the plots will not necessarily be in the order expected. When I ran this with a regular HASH, the graphic came out as "squared" on the left, "linear" in the middle, and "cubed" on the right.

Finally, I used a list to hold the object returned by plot,which enabled me to change the color of the plots after the graphic initialized. The LIST::ADD method adds the plots to the list one at a time.

Please login or register to post comments.