The SWC-DB versions 0.4.14 and 0.4.15 introduce the new Column Type SERIAL and Domain-Object Conditions with new Comparators subset, superset, partially ordered subset/superset and fully ordered subset/superset, Plus the option of Multiple Value Expressions.
The new features, As Example, in Audio Recognition can be used in the following manner.
The Column Type of the schema is SERIAL and Key Sequence is VOLUME.
The KEY is set with Fractions of normalized audio with pairs of Frequency with Absolute Milli-Amplitude plus the corresponding Audio ID and the second of the timeline:
[10000, 100, 5000, 200, 2500, 150, 1250, 300, 625, 200, 300, 100, 150, 300, 50, 444, AUDIO#, SECOND]
and the VALUE is set with a LIST_INT64 of samples (on base of uA INT for the reason Double is more complex in computing) that correspond to the Second of the Audio# Timeline. [ ..., 1000000, 80000, 80010, -10000, -20000, 50000, ... ], It is a matter of required precision and the sample-rate reduction to save on Query Time and Storage space.
The query corresponding to the dataset can be expressed with SWC-DB SQL as the follow query:
select col(TheAudiosColumnName) = (cells=(
[10000, 90, 5000, 190, 2500, 140, 1240, 300, 625, 190, 300, 90, 150, 290, 50, 434, =>""]
<= KEY <=
[10000, 110, 5000, 210, 2500, 160, 1260, 300, 625, 210, 300, 110, 150, 310, 50, 454, =>""]
AND
VALUE = [0:LI:->[ ..., 1000000, 80000, 80010, -10000, -20000, 50000, ... ]]
ONLY_KEYS
) DISPLAY_STATS );
The description behind the query command, It is Select cells with key matching -/+10mA of all Normalized Absolute Frequencies and those that match Partially-Ordered Subset and The result is the corresponding Cells Key with Audio IDs Fraction that is in-question for the given audio-samples.
For a close match (not-exact) within the -/+10000uA, the Query can be extended to:
AND
VALUE = [0:LI:->[ ..., >=990000, >=70000, >=70010, =>-20000, =>-30000, =>40000, ... ]]
AND
VALUE = [0:LI:->[ ..., <=
1000000 , <=90000, <=90010, <=0, <=-10000, <=60000, ... ]]
Related SWC-DB Documentations References: