In-Memory SQL-Like Database

Build an in-memory database that supports creating tables, inserting integer rows, and selecting rows with AND-combined predicates over comparison operators. Requirements Input is list[list[str]], where each inner list represents a tokenized query. Support at least these operations: create table <tablename> ( <col1> <col2> <col3> ) insert into <table> ( <col_val1> <col_val2> <col_val3> ) select from <tablename> where ( <fieldA> <op1> <valueA> AND <fieldB> <op2> <valueB> ) Constraints and assumptions: All column values are integers. Selection operators include >=, =, <=, and similar comparisons. select only needs to support AND between predicates. Query format is assumed valid, but field values may be invalid and should be handled deliberately. Minimum goal is a working end-to-end implementation; polished validation can come after the main path works. Notes Keep schema metadata separate from row storage: table_name -> columns, table_name -> rows. Parse first for operation type, then dispatch. Avoid over-engineering a full SQL parser; the prompt uses constrained token lists. For select, map column names to indices and evaluate each predicate against the row's integer value. Candidate experience emphasizes speed: parsing plus AND support can consume the round if the basic path is not implemented first. Preparation Write a constrained SQL-token interpreter with create, insert, and select in one sitting. Drill predicate evaluation for =, <, <=, >, >= and invalid column/value handling. Prepare a simple table/row data model before the interview so parsing does not dominate.