How does indexing affect database size


(Folterj) #1

It's important to be able to accurately estimate the size of datebases at time of design, especially if scaling is involved.

This article describes the basic size calculations, and also includes an automated calculator:

What's missing is indexing, and how this affects the size. There is no mention of this anywhere on the website in fact.
Observations from some simple testing with 1M nodes:

  • Adding 1 or even 2 unique UUID properties on each node does not increase the size at all (contrary to documentation and calculator reporting 41 bytes per property)
  • Indexing (the new v3.5 indexing format) on UUID greatly increases the database size: around 72 bytes per node (e.g. 1M nodes + UUID property; total size = 14MB, add indexing; total size = 86MB)

Can you confirm or comment on this please?
Suggest indexing to be included in documentation, and in the online calculator please.