Hello all
While trying to model my data to use OrientDB's Document database, I
am conflicted between employing two alternative strategies in trying to
fasten read times. Penalty due to higher read times is quite large and
hence my concern.
1. In the traditional approach, employ indexing on this table and query for
a combination of columns such as userID, date, type of record to narrow
down to one particular record. Ideally, I should have architected the model
to employ least such combination but due to reasons such as simplicity and
resource constraints, I had to resort to this style. If I were to use this,
I could employ a composite index for this combination and fetch records.
2. Alternatively, I was wondering if I can make use of Record ID mapping.
Let's say, if I can manage to architect a mapping that can help me navigate
my way through the search combination (UserID >> date >> type) using a
separate document class(es), I might be able to reach the record
straightaway by navigating this path. Of course, I will have to create and
retain map for each record that gets created, but I guess I can live that
if my read times are better using this approach. I can additionally
provision CRC if deemed necessary.
Has anybody been through this path? If yes, what was the overall
experience? I employ JAVA API for record creation and fetching.
thanks
Chandra
--
---
You received this message because you are subscribed to the Google Groups
"OrientDB" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
For more options, visit https://groups.google.com/d/optout.
