Struct QueryBuilder 

pub struct QueryBuilder;

Query builder for converting plans to SQL.

Implementations

impl QueryBuilder

pub fn build_read(tree: &ReadPlanTree) -> Result<SqlFragment>

Build a SELECT query from a read plan tree.

pub fn column_sql(field: &CoercibleField) -> String

A column reference with its JSON path, as it appears in SQL.

Exposed for ordering by an embedded resource’s column, which is assembled outside the read plan.

pub fn qualified_column_sql( qualifier: Option<&str>, field: &CoercibleField, ) -> String

The same, with the column qualified by a relation alias.

The alias belongs to the column, not to the expression around it: to_jsonb(e2."settings")->'a', never e2.to_jsonb(...), which reads as a function in a schema called e2.

pub fn filter_sql(filter: &Filter, pg_type: &str) -> Result<SqlFragment>

Build the SQL for a single filter against a column of pg_type.

Exposed for the embedding path, which applies filters inside a child subquery it assembles itself rather than through a full plan. Column names are left unqualified: inside the child’s subselect the innermost FROM wins, so a bare name binds to the child. The caller is responsible for having checked the column exists there – otherwise the name would resolve outward to the correlated parent instead.

pub fn logic_sql<F>(tree: &LogicTree, type_resolver: F) -> Result<SqlFragment>

where F: Fn(&str) -> String + Copy,

Build the SQL for a logic tree against a table whose column types type_resolver supplies.

Exposed for the same reason as filter_sql: an embedded resource’s and=/or= is applied inside a subquery the embedding path assembles itself, without going through a full plan.

pub fn build_mutate(plan: &MutatePlan) -> Result<SqlFragment>

Build a mutation query.

pub fn build_call( plan: &CallPlan, read: Option<&ReadPlanTree>, ) -> Result<SqlFragment>

Build an RPC call query.