ArcadeDB supports pagination natively. Pagination doesn’t consume server side resources because no cursors are used. Only Record ID’s are used as pointers to the physical position in the bucket.
There are 2 ways to achieve pagination:
The first and simpler way to do pagination is to use the SKIP/LIMIT approach. This is the slower way because ArcadeDB repeats the query and just skips the first X records from the result.
Syntax:
SELECT FROM <target> [WHERE ...] SKIP <records-to-skip> LIMIT <max-records>Where: - records-to-skip is the number of records to skip before starting to collect them as the result set - max-records is the maximum number of records returned by the query
This method is faster than the SKIP-LIMIT because ArcadeDB will begin the scan from the starting RID. ArcadeDB can seek the first record in about O(1) time. The downside is that it’s more complex to use.
The trick here is to execute the query multiple times setting the LIMIT as the page size and using the greater than > operator against @rid. The lower-rid is the starting point to search, for example #10:300.
Syntax:
SELECT FROM <target> WHERE @rid > <lower-rid> ... [LIMIT <max-records>]Where: - lower-rid is the exclusive lower bound of the range as RID - max-records is the maximum number of records returned by the query
In this way, ArcadeDB will start to scan the bucket from the given position lower-rid + 1. After the first call, the lower-rid will be the rid of the last record returned by the previous call. To scan the cluster from the beginning, use #-1:-1 as lower-rid .