Add ability to limit number of results about cqengine HOT 8 CLOSED

Actually what you probably need is a NavigableIndex on price (i.e. not a HashIndex), and to request CQEngine to use the "index" ordering strategy. Take a look at: OrderingStrategies.md

If you need more help with this, feel free to open a new issue and we can continue discussing there. Hope that helps for now!

from cqengine.

Original comment by [email protected] on 28 Aug 2013 at 4:57

• Added labels: Type-Enhancement
• Removed labels: Type-Defect

from cqengine.

Pagination would be a common requirement for large collection. In addition to 
sequential paging as shown in the example, arbitrary start positions should be 
supported: Pagination(results, 20, 20) [ start at position 20 (1 based) and 
return 20 records).
Guava's Iterable and FluentIterable interfaces support such pagination and 
could be used to wrap a ResultSet.

Original comment by [email protected] on 22 Jan 2015 at 9:51

from cqengine.

Closing this issue as this problem is best solved by combining other libraries with CQEngine.

CQEngine's ResultSet implements the java.lang.Iterable interface, therefore results can be limited or paginated as follows:

• Guava's Iterables.limit()
• Convert the CQEngine ResultSet into a Java 8 Stream and call Stream.limit()
• Consider using a variety of open source pagination libraries.

from cqengine.

Please open this issue. Here is the reasoning:

Records: 100k

SQL:
SELECT * FROM Cars ORDER BY price LIMIT 10

cqengine will sort and return all 100k records (2+ seconds), and then other libraries should paginate.

If SQL parser considers LIMIT keyword and finds defined index (HashIndex on price), it would be natural to optimize response to top 10 results by HashIndex.

As a result, query will be in a ms instead of sec.

from cqengine.

Create a limitBy option, add it to the queryOptions. Create your own version of MaterializedOrderedResultSet that only sorts the top n of the resultSet and just appends the rest to the bottom of the data structure. Upon iterator(), read that limit and sort accordingly. This way you'll have a "list" of resultset.size where the top n are sorted (given that you don't care about n+1...) rather than using the treeSet to sort the entire result set. Any item that compares > to the nth item is just added to the bottom of the semi-sorted structure. You could also use a variation of the binary search to reduce the complexity of finding where to insert an item in the top n. Try it out by sloppily overriding MaterializedOrderedResultSet in your project path with hard-coded limits, if it works out for you build your own queryEngine etc

from cqengine.

great work btw @ npgall

from cqengine.

please someone give me a example how to achieve pagination

from cqengine.