.. default-domain:: chpl

.. module:: Unique
   :synopsis: unique finding and counting algorithms

Unique
======
**Usage**

.. code-block:: chapel

   use Unique;


or

.. code-block:: chapel

   import Unique;

unique finding and counting algorithms
these are all based on dense histograms and sparse histograms(assoc domains/arrays)

you could also use a sort if you got into a real bind with really
large dense ranges of values and large arrays...

*** need to factor in sparsity estimation somehow ***
for example if (a.max-a.min > a.size) means that a's are sparse



.. data:: const uLogger = new Logger(logLevel, logChannel)

.. function:: proc uniqueSort(a: [?aD] ?eltType, param needCounts = true) throws

   
   sorting based unique finding procedure
   
   Returns a tuple: (UniqueValArray,UniqueValCountsArray)
   which contains the unique values of a, along with the number of times each unique value appears in a
   
   :arg a: Array of data to be processed
   :type a: [] int
   
   :returns: ([] int, [] int)
   

.. function:: proc uniqueSortWithInverse(a: [?aD] ?eltType, param needIndices = false) throws

.. function:: proc uniqueFromSorted(sorted: [?aD] ?eltType, param needCounts = true) throws

.. function:: proc uniqueGroup(str: SegString, returnInverse = false) throws

.. function:: proc uniqueFromTruth(str: SegString, perm: [?aD] int, truth: [aD] bool) throws