\ Aggregate Btree                                                    vandys
only extensions definitions

AgView -> subclass: AgBtree





















\ AgBtree fetch                                                      vandys

AgBtree -> :method @ ( key self -- val T | F )   ag>list { key list }
   list -> size   0   do   i list -> @ { bt }
      key bt -> @ if exit then   loop false method;




















\ AgBtree iteration                                                  vandys

AgBtree -> :method do { arg 'fn self -- }
   0 begin   self -> nextKey   while { key }
      key self -> @ drop { val }   arg val key 'fn execute
   repeat method;



















\ Aggregate B-trees, testing                                         vandys
false [if] \ Testing

Btree -> new constant bt1   Btree -> new constant bt2
AgBtree -> new constant ab

100 1 bt1 -> !   200 2 bt2 -> !   300 3 bt1 -> !   300 3 bt2 -> !

bt1 ab -> add   bt2 ab -> add

[then] \ Testing

AgBtree is an AgView of Btree's























:method @ Return value for this key from whichever B-tree first has it
























:method do Iterate across all values in the B-tree.  If a value is at
 a given key in more than one layer, only the outermost one is seen
 by the iterator function.
TBD: if this goes too slow, may need to maintain parallel indices
 into the constituent B-trees rather than iteratively re-indexing
 into each of them to find successive steps.