The big { } code block without the END gets run for every line. The values of $1, $2, $3, etc are set to the values for the columns read. Putting variables and strings in a row sticks them together into a longer string.

So every line, it sets a value like this in the array:

Code:

T["a b c"]="d e f"

If there's a duplicate line, setting the same value in the array twice doesn't put two elements in the array. The previous contents of T["a b c"], for instance, would just get overwritten given another line starting with a b c

Once all the lines have been read, only then will awk run the END {} block, which goes through each thing in the array and prints them (in no particular order).
The syntax for(X in ARRAY) loops through every element in an array, with X being the array index, and ARRAY[X] being the contents of that index.