Database Compactness?

[1/2] from: robbo1mark::aol::com at: 21-May-2002 9:43

This is one for the guru's or Rt staff, Which of the following provides a more compact and secure representation for a database? an object! db1: context [ group1: [#1 "A Value"] ] or a block! db2: reduce [ 'group1 [#1 "A Value"] ] They both allow path retrieval ie.

>> db1/group1/#1

== "A Value"

>> db2/group1/#1

== "A Value"

As well as insertions / removals

>> append db1/group1 [#2 "A Thing"]

== [#1 "A Value" #2 "A Thing"]

>> probe db1

make object! [ group1: [#1 "A Value" #2 "A Thing"] ]

>> append db2/group1 [#2 "A Thing"]

== [#1 "A Value" #2 "A Thing"]

>> probe db2

[group1 [#1 "A Value" #2 "A Thing"]] I welcome help / opinions on this matter for the REBOL-TRIPLES project to design and create REBOL/ASSOCIATES database which aims to produce a "Sentences" like product with associated model database. thanx, Mark Dickson

[2/2] from: ingo:2b1 at: 21-May-2002 23:49

Hi Mark, [Robbo1Mark--aol--com] wrote:

> This is one for the guru's or Rt staff, > > Which of the following provides a more compact and secure representation for a database? > > an object! > > db1: context [ group1: [#1 "A Value"] ] > > or a block! > > db2: reduce [ 'group1 [#1 "A Value"] ]

I don't know about the security aspect, but think they'd be equal. Comparing creation time / space efficency, blocks clearly win I'd _guess_ that objects are faster for lookups, but I don't know. (And using hashes instead of blocks is the worst for creation time / space, but I don't know how it works for lookups.)

> They both allow path retrieval ie. > >>>db1/group1/#1 >> > == "A Value"

<..>

> As well as insertions / removals > >>>append db1/group1 [#2 "A Thing"] >> > == [#1 "A Value" #2 "A Thing"]

<..> But there would be a problem when it comes to adding a new group to the object! case, I don't know if this might pose a problem in this specific case, though. I hope that helps, Ingo