Lesson 8: Structuring query results

Clauses can be chained arbitrarily into query pipelines, since they all consume streams of concept rows and produce stream of concept rows

match <pattern>
reduce <operation>
match <pattern>
sort <vars>;
insert <pattern>
delete <pattern>
match <pattern>
limit N;
...

fetch clauses can only be placed at the end of a query pipeline to transform the stream into a stream of JSON documents

Use reduce operators to aggregate streams

The reduce operators are count, sum, max, min, mean, median, and std. Except for count, each modifier takes a variable to aggregate over as an argument, which can represent either a numeric attribute or a numeric value.

You can group reductions by using groupby

reduce $count = count groupy $book;

The query streams can be sorted using the sort operator

match <pattern>
sort var;

The results of queries can be paginated using the limit and offset modifiers, which take a literal integer as an argument.

match <pattern>
offset 10;
limit 5;

The sort, limit, and offset modifiers can all be used individually or in any combination.

Subqueries can be encapsulated and invoked easily wth functions.

Fetch clauses can project attributes

fetch {
 "value": $owner.attribute-type,
};

Attributes, values, and types can be fetched directly in the fetch clause

fetch {
  "attribute": $attribute,
  "type": $type_,
  "computed-value": $computed-value,
};

Subqueries can be written directly into the fetch clause with the correct encapsulation:

Underlying idea: all leaves in a fetch clause structure must be serializeable to a value