A little toolkit for multiset (bag) basic operations, storing bags into JSONb objects.

Two flavors: schema lib implementation and direct implementation (schema public).

As a restric jsonb-schema, with only key-value pairs and value as integer, the datatype is named jbags.

• bag.valid(JSONb) ret boolean:
  Validate internal representation as bag (key-multiplicity pairs).
• bag.sc(JSONb) ret bigint:
  Scalar cardinality of the bag.
• bag.contains(JSONb, text) ret boolean:
  Checks whether the element is present (at least once) in the bag.
• bag.j_as_t(JSONb[]) ret SETOF bag.ref:
  Returns an array of JSONb-bag representations as an element-multiplicity table.
• bag.j_as_t(JSONb) ret SETOF bag.ref:
  Similar.
• bag.scaled_by(JSONb,int) ret JSONb:
  Scalar multiplication. $2⊗$1.
• bag.intersection(JSONb[]) ret JSONb:
  ...
• bag.union(JSONb[]) ret JSONb:
  ...
• bag.sum(JSONb[]) ret JSONb:
  ...
• bag.is_sub(JSONb, JSONb[]) ret boolean:
  Checks $1 ⊑ ($2[1] ∩ $2[2] ∩...), that is if $1=($1 ∩ $2[1] ∩ $2[2] ∩...).
• bag.intersection(JSONb,JSONb) ret JSONb:
  $1 ∩ $2.
• bag.is_sub(JSONb, JSONb) ret boolean:
  Checks $1 ⊑ $2, whether each element in the bag1 occurs in bag1 no more often than it occurs in the bag2.
• bag.union(JSONb,JSONb,boolean DEFAULT true) ret JSONb:
  ... Trying to optimize the union (or sum) of two bags.
• bag.j_as_t(JSONb) ret SETOF bag.ref:
  

See bag_as_schema.sql source code.

Same as bag.*() but in the public catalog of functions with jbag datatype as parameter.
NOTE: j_as_t() renamed to astable().

See bag.sql source code.