Pcount

Aggregation functions pcount

syntax

• pcount(a)
• pcount_uint8(a), pcount_uint16(a), pcount_uint32(a), pcount_uint64(a)

description

pcount is short for partition count, and is used to obtain a count for each possible value of a relational attribute, i.e. an attribute with a values unit that is or can be the domain of other attributes.

The domain of the results of these functions is set to the values unit of the argument a.

• pcount(a) is defined as the count of the number of entries of a relation in the related domain unit. The values type is the cardinality type of the values of the domain of a:
  • Int8, UInt8 -> UInt8;
  • Int16, UInt16 -> UInt16;
  • Int32, UInt32, SPoint, WPoint -> UInt32;
  • Int64, UInt64, IPoint, UPoint -> UInt64.
• pcount_uint(8|16|32|64)(a) are pcount variants resulting in uint8, uint16, uint32, or uint64 data items respectively, which saves memory when one knows that the number of occurrences of each value of a does not exceed the maximum number of the specified values type.

Use the has_any (a) function if only the occurrence of values is required.

applies to

• attribute a with uint2, uint4, uint8, uint16, uint32, uint64, spoint, wpoint, ipoint, upoint or bool value type

performance

O(n) where n = number of elements in the domain of a. Single pass counting occurrences of each value.

example

attribute<uint32> pcountReg (Region) := pcount(City/Region_rel);

City/Region_rel
0
1
2
1
3
null
3

domain City, nr of rows = 7

pcountReg
1
2
1
2
0

domain Region, nr of rows = 5

see also

• count
• has_any