Characterisation bitfields description

The following table specifies how the ResultField values for

ResultField.POSTCODE_CHARACTERISTICS
ResultField.TOWN_CHARACTERISTICS
ResultField.STREET_CHARACTERISTICS
ResultField.HOUSENR_CHARACTERISTICS

can be interpreted. The table shows a list of different characteristics. Each characteristic is associated with a number, which is given in decimal and hexadecimal format. The return value for one of the four ResultField types is a unique sum of these decimal (or hex) numbers.

Example

Here's an example:
Assume the return value is 4996. The sum for 4996 is 4996 = 4096 + 512 + 256 + 128 + 4. Therefore the characteristics for this value are

Result words include input (4096)
Input is prefix of phonetic Multiword result (512)
Input is prefix of Multiword result (256)
Input is phonetic prefix of result (128)
Input is prefix of result (4)

How do you find the unique sum for a value?
A quite simple way is to repeatedly subtract the highest possible number from the input value until you reach zero. So if you start with 4996 then 4096 is the highest value that can be subtracted. The rest is 900 and thus 512 is next number that can be subtracted. Next is 256 and so on.

The value can also be decoded using the hex value. The decimal value 4996 is equal to the hex value 0x00000384, which is the sum of

0x00000004
+0x00000080
+0x00000100
+0x00000200
+0x00001000

The characterisation table

First some definitions...

The columns postcode, town, street and housenr correspond to the enum values ResultField.POSTCODE_CHARACTERISTICS, ResultField.TOWN_CHARACTERISTICS, ResultField.STREET_CHARACTERISTICS and ResultField.HOUSENR_CHARACTERISTICS.
The icons show the availability of each characteristic on the four ResultField types.
The steps listed in the column Preprocessing are applied in the given order on both input and result. Then, the characteristic is true if the preprocessed data fulfils the definition. Following preprocessing steps are available:
- SN Special character normalisation: Special characters and umlauts are replaced (e.g. ü → ue).
- CN Case normalisation: Upper and lower case characters are treated in the same way.
- AN Abbreviation normalisation: The replacements defined in the abbreviation dictionary are applied.
- XN Affix normalisation: All affixes defined in the affix dictionary are removed from the string.
- PN Phonetic normalisation: The phonetic replacements are applied.
- SepN Separator normalisation: All separators (Space, -, /, ., etc.) are treated in the same way. Multiple consecutive separators are projected to a single one.
Prefix: x is Prefix of another word y if x is not longer than y and characters on position i (i = 1 ... length( x )) in x and y are identical.
Multiword Prefix: A list of words w is a Multiword Prefix of a list of words z, if an injective mapping from the words w₁ ... w_n to the words z₁ ... z_m exists such that each word w₁ ... w_n is Prefix of its corresponding word in z₁ ... z_m. (Note that m >= n)

Characteristic	Availability				Preprocessing	Definition	Examples
	postcode	town	street	housenr			Positive	Negative
Input is prefix of result Decimal 4 - Hex 0x00000004					CN, SN, SepN	Input is Prefix of result	I: karl O: Karlsruhe	I: karle O: Karlsruhe
Input is phonetic prefix of result Decimal 128 - Hex 0x00000080					CN, SN, SepN, PN	Input is Prefix of result	I: karl O: Karlsruhe I: carl O: Karlsruhe	I: karle O: Karlsruhe
Input is prefix of Multiword result Decimal 256 - Hex 0x00000100					CN, SN, SepN	Input is Multiword Prefix of result	I:Wei Beu O:Beutelsbach Weihersdorf I: Beu O:Beutelsbach Weihersdorf	I:BeuWei O: Beutelsbach Weihersdorf I: Beu Wei O:Beutelsbach
Input is prefix of phonetic Multiword result Decimal 512 - Hex 0x00000200					CN, SN, SepN, PN	Input is Multiword Prefix of result	I:Karls Turla O:Karlruhe Durlach I:Carls O:Karlruhe Durlach	I:BeuWei O: Beutelsbach Weihersdorf I: Beu Wei O:Beutelsbach
Match except abbreviations Decimal 1024 - Hex 0x00000400					CN, SN, SepN, AN	Input is Multiword Prefix of result	I:Stumpfstrasse O:Meine Stumpfstr	I:Stumpfstrasse O:Stumpfweg I:Stumpfstrasse O:Stampfstr
Match except affixes Decimal 2048 - Hex 0x00000800					CN, SN, SepN, XN	Input is Multiword Prefix of result	I:Stumpfweg O:Stumpfstr	I: Stumpfstra O: Stumpfstr
Result words include input Decimal 4096 - Hex 0x00001000					CN, SN, SepN, AN	Input is Multiword Prefix of result	I: Beu O:Beutelsbach Weihersdorf	I: Beu Wei O:Beutelsbach
Input words include result Decimal 8192 - Hex 0x00002000					CN, SN, SepN, AN	Result is Multiword Prefix of input	I: Beutelsbach Wei O:Beutelsbach	I: Beu O:Beutelsbach Weihersdorf
Match representative postal code Decimal 16384 - Hex 0x00004000					CN, SepN	The representative postal code matches to the input	I:76000 O:76* I:76000 O:76229, RepPostcode: 76*	I:76000 O:75000
Match length Decimal 8 - Hex 0x00000008					CN, SN, SepN	The length of the input and the result is the same	I:Beutelsbach - Weihersdorf O: Beutelsbach Weihersdorf I: Karlsruhe O: Stuttgart	I:Beu Wei O: Beutelsbach Weihersdorf
Match exact Decimal 1 - Hex 0x00000001					CN, SN, SepN, AN	Result is Multiword Prefix of input and Input is Multiword Prefix of result	I: Weihersdorf Beutelsbach O: Beutelsbach Weihersdorf	I: Weihersdorf Beutelsdorf O: Beutelsbach Weihersdorf
Match exact Decimal 1 - Hex 0x00000001					CN, SepN	The strings are identical	I: ab3-b5 O: AB3 B5	I: AB3B5 O: AB3 B5
Match exact Decimal 1 - Hex 0x00000001					(none)	The strings are identical	I: 42 O: 42	I: 42a O: 42
Match except separators Decimal 16 - Hex 0x00000010					(delegated to Match exact)	After removing all separators Match exact applies	I:Media Park O:Mediapark	I:Media Par O:Mediapark
Match quality high Decimal 32 - Hex 0x00000020					(various)	The rating algorithm of the gpGeocoder indicates a high quality	I: Stampfstr O: Stumpfstr	I: Bbuch O: Beutelsbach Weihersdorf
Match quality medium Decimal 64 - Hex 0x00000040					(various)	The rating algorithm of the gpGeocoder indicates a medium quality	I: Stamfstr O: Stumpfstr	I: Karlsruhe O: Berlin
Match quality low Decimal 2 - Hex 0x00000002					(various)	The rating algorithm of the gpGeocoder indicates a low quality	I: Karlsruhe O: Berlin	I: Karlsruhe I: Stamfstr O: Stumpfstr
CityFieldPermutation Decimal 32768 - Hex 0x00008000					CN, SN, SepN, AN	Not(Considering input City field and result City field Input is Multiword Prefix of result is fulfilled. The same condition must hold for city2.)	I: C: Durlach O:C: Karlsruhe C2: Durlach	I: C: C2: Durlach O: C: Karlsruhe C2: Durlach
Match exact except additions Decimal 65536 - Hex 0x00010000					(delegated to Match exact)	If additions are removed, input and result Match exact	I: 42a O: 42	I: 4 2a O: 42