Onsite Coding: Expression Evaluation with Variables

A recurring onsite coding prompt combining calculator parsing with variable assignment / equation evaluation. Examples include `A = 5`, `B = A`, `C = 5`, `D = C - 10`; variants ask for DFS over variable dependencies. Requirements Input is a set of variable definitions or equations plus arithmetic expressions. Variables can point to numeric literals or to other variables. Expressions include simple + and -; some variants are close to basic calculator and may include parentheses. Return the evaluated value for a requested variable or expression. Detect undefined variables and cycles if the interviewer asks for robustness. Example shape: A = 5 B = A C = 5 D = C - 10 query D -> -5 DFS variant: equations = [["a", "b"], ["b", "c"]] values = [2.0, 3.0] query a / c -> 6.0 Notes Split the problem into parsing and evaluation. Store each variable as an expression AST or a token list; then evaluate with memoized DFS. For simple +/-, a stack or running sign * value accumulator is enough. Parentheses require recursive descent or the standard stack-based Basic Calculator pattern. For equation-division variants, build a weighted graph and run DFS/BFS from numerator to denominator, multiplying edge weights along the path. Cycle handling: maintain visiting and visited sets. If a variable is encountered while already visiting, return an error / unknown value rather than recursing forever. Complexity: parsing is O(total input length). Evaluation is O(V + E) for graph/equation variants and O(length) per expression with memoization over variables. Preparation Drill Basic Calculator, Evaluate Division, and a tiny variable-substitution parser. Practice explaining why parsing into tokens first is safer than doing repeated string replacement. Prepare tests for chained variables, undefined variables, a negative result, and a cycle.