Why are ranges split into multiple?

[rynop]

When generating the queries to dispatch to DDB, this library goes through all the cells in a covering area, and splits the low and hi range of each cell into sub-ranges based on the hash key length.

I've gone through the logic quite a few times and I can't figure why trySplit() (splitting of cell range into multiple) is needed.

Can't I just take the lowest minimum range of all the cells, and the highest maximum range of all the cells, and construct a DDB BETWEEN condition on the local secondary index?

For example, if my covering had 3 cells and I have a hashKeyLength of 3:

cell 1:
  - rangeMin:    123456789
  - rangeMax    123999999

cell 2:
  - rangeMin:    124000000
  - rangeMax    124999999

cell 3:
  - rangeMin:    125000000
  - rangeMax    125678912

Can't I just do BETWEEN 123456789 and 125678912?

Why do I have to trySplit() cell1s range (123456789-123999999) into 3 separate ranges?