View source code
Display the source code in std/range/package.d from which this page was generated on github.
Report a bug
If you spot a problem with this page, click here to create a Bugzilla issue.
Improve this page
Quickly fork, edit online, and submit a pull request for this page. Requires a signed-in GitHub account. This works well for small changes. If you'd like to make larger changes you may want to consider using local clone.

Enum std.range.SearchPolicy

Policy used with the searching primitives lowerBound, upperBound, and equalRange of SortedRange below.

The enum base type is int .

Enum members

Name Description
binarySearch Searches using a classic interval halving policy. The search starts in the middle of the range, and each search step cuts the range in half. This policy finds a value in Ο(log(n)) time but is less cache friendly than gallop for large ranges. The binarySearch policy is used as the last step of trot, gallop, trotBackwards, and gallopBackwards strategies.
gallop Performs a galloping search algorithm, i.e. searches with a step that doubles every time, (1, 2, 4, 8, ...) leading to an exponential search schedule (indexes tried are 0, 1, 3, 7, 15, 31, 63,...) Once the search overshoots its target, the remaining interval is searched using binary search. A value is found in Ο(log(n)) time.
gallopBackwards Similar to gallop but starts backwards. Use it when confident that the value is around the end of the range.
linear Searches in a linear fashion.
trot Searches with a step that is grows linearly (1, 2, 3,...) leading to a quadratic search schedule (indexes tried are 0, 1, 3, 6, 10, 15, 21, 28,...) Once the search overshoots its target, the remaining interval is searched using binary search. The search is completed in Ο(sqrt(n)) time. Use it when you are reasonably confident that the value is around the beginning of the range.
trotBackwards Similar to trot but starts backwards. Use it when confident that the value is around the end of the range.

Authors

Andrei Alexandrescu, David Simcha, Jonathan M Davis, and Jack Stouffer. Credit for some of the ideas in building this module goes to Leonardo Maffi.

License

Boost License 1.0.

Comments