06 Operators

06 FQL Operators

AF: attribute function

Filters

signature	parameters	input function	output function	semantic	example
filter_items	predicate (phrased against items)	any AF	new instance, same type as input function	returns a new instance of the input function containing only the items qualifying under the given predicate	filter relation function on conjunction of its key and its value
filter_values	predicate (phrased against values)	any AF	new instance, same type as input function	same as filter_items, but predicate phrased against values	filter values of a relation function
filter_keys	predicate (phrased against keys)	any AF	new instance, same type as input function	same as filter_items, but predicate phrased against keys	filter keys (i.e. relation function names) of a database function

Transforms

signature	parameters	input function	output function	semantic	example
transform	transformation_function (phrased against items)	any AF	new instance as created by transformation_function	returns a new transformed instance, similar to existing map() in Python, but a valid FQL operator which may be pushed down to an engine	project input AF, return a computed AF based on input AF, project
transform_items	transformation_function (phrased against items), output_factory (for the attribute function to return)	any AF	new instance as created by the output factory	returns a new instance of the input function containing the transformed items of the input AF	aggregate multiple items into a single item (like done in classical aggregation), project items of input AF
partition	partitioning_function (phrased against items), output_factory (for the attribute function to return)	any AF	new instance as created by the output factory	returns a new instance of the input function containing the partitions	horizontally partition (aka grouping) or vertically partition the input AF, replicate the input AF, any combination of the latter
group_by	partitioning keys (phrased against items), output_factory (for the attribute function to return)	any AF	new instance as created by the output factory	returns a new instance of the input function containing the partitions	horizontally partition (aka grouping) or vertically partition the input AF, replicate the input AF, any combination of the latter
aggregate	aggregate input AF based on the specified aggregation functions	any AF	same instance type as input AF	returns a new instance of the input function containing the aggregation results	classical aggregation without grouping but including schema definition of the created AF
group_by_aggregate	partitions input AF based on a lambda function, then aggregates each input partition into an output AF	any AF	new instance as created by the output factory	returns a new instance of the input function containing the transformed items	classical group_by_aggregate (which in SQL is three operators: grouping, aggregation, projection)

Subdatabase Operators

signature	parameters	input function	output function	semantic	example
subdatabase	TODO	DBF	DBF	returns a DBF reduced to the participating TFs	classical Yannakakis-reduction, resultdb
grouping_set	TODO	DBF	DBF	returns a DBF with grouping sets	classical grouping sets, but not hacked into one output relation as in SQL
cube	TODO	DBF	DBF	returns a DBF with cube results	classical cube, but not hacked into one output relation as in SQL
outer	TODO	DBF	DBF	returns a DBF with additional outer partitions	subdatabase plus outer partitions for the RFs specified, but not hacked into one output relation as in SQL
anti	TODO	DBF	DBF	returns a DBF with additional anti partitions	subdatabase plus anti partitions for the RFs specified, but not hacked into one output relation as in SQL

(Flattening) Joins

signature	parameters	input function	output function	semantic	example
join	TODO	DBF	RF	returns a join RF of the input DBF	classical join
equi_join	TODO	DBF	RF	returns a join RF of the input DBF	classical join

TODO, ongoing...