sqlode - v0.15.0

sqlode/query_analyzer/param_inferencer

Values

extract_int_context_params

</> 
pub fn extract_int_context_params(
  tokens: List(lexer.Token),
  engine: model.Engine,
) -> dict.Dict(Int, model.ScalarType)

Extract placeholder type hints from SQL contexts that the spec requires to be integer: the expressions following LIMIT and OFFSET. Each placeholder reachable from those keywords (until a terminating keyword or end of statement) is pinned to IntType. Used as a complement to extract_type_casts so callers can write LIMIT sqlode.arg(lim) / OFFSET sqlode.arg(off) without an explicit cast on engines like SQLite that cannot infer the type from a bare ? (Issue #491).

Implementation note: this works at the token level rather than the IR, because the IR’s Param.index is 0 for SQLite anonymous ? placeholders (expr_parser.decode_placeholder deliberately returns 0 for them). The token walker assigns each placeholder an occurrence index that matches what placeholder.build computes, so the resulting dict keys line up with PlaceholderOccurrence.index consumed by build_params.

extract_type_casts

</> 
pub fn extract_type_casts(
  ctx: context.AnalyzerContext,
  engine: model.Engine,
  tokens: List(lexer.Token),
) -> Result(dict.Dict(Int, model.ScalarType), #(Int, String))

infer_equality_params

</> 
pub fn infer_equality_params(
  ctx: context.AnalyzerContext,
  engine: model.Engine,
  query_name: String,
  tokens: List(lexer.Token),
  catalog: model.Catalog,
) -> Result(List(#(Int, model.Column)), context.AnalysisError)

infer_in_params

</> 
pub fn infer_in_params(
  ctx: context.AnalyzerContext,
  engine: model.Engine,
  query_name: String,
  tokens: List(lexer.Token),
  catalog: model.Catalog,
) -> Result(List(#(Int, model.Column)), context.AnalysisError)

infer_insert_params

</> 
pub fn infer_insert_params(
  ctx: context.AnalyzerContext,
  engine: model.Engine,
  tokens: List(lexer.Token),
  catalog: model.Catalog,
) -> List(#(Int, model.Column))

infer_insert_params_from_ir

</> 
pub fn infer_insert_params_from_ir(
  engine: model.Engine,
  statement: query_ir.SqlStatement,
  catalog: model.Catalog,
) -> List(#(Int, model.Column))

Structured IR variant of infer_insert_params. Consumes the pre-parsed InsertStatement directly instead of re-scanning the token list.

✨ Search Document