# spreadsheetbench-verified / 38074 - taskset: [spreadsheetbench-verified](https://harnessreport.com/tasks/spreadsheetbench-verified.md) - difficulty: hard - category: spreadsheet-manipulation - language: - runnable from the site: no - agent timeout: 600s ## Results by harness _none yet_ ## Instruction ``` You are a spreadsheet expert who can manipulate spreadsheets through Python code. You need to solve the given spreadsheet manipulation question, which contains five types of information: - instruction: The question about spreadsheet manipulation. - spreadsheet_path: The path of the spreadsheet file you need to manipulate. - instruction_type: There are two values (Cell-Level Manipulation, Sheet-Level Manipulation) used to indicate whether the answer to this question applies only to specific cells or to the entire worksheet. - answer_position: The position need to be modified or filled. For Cell-Level Manipulation questions, this field is filled with the cell position; for Sheet-Level Manipulation, it is the maximum range of cells you need to modify. You only need to modify or fill in values within the cell range specified by answer_position. - output_path: You need to generate the modified spreadsheet file in this new path. Below is the spreadsheet manipulation question you need to solve: ### instruction I need to calculate the profit in a table by subtracting expenses (1,2,3). The columns contain both profit and expenses values, but I need to evaluate this calculation based where column A matches the case sensitive values in J2 using vectorized operations. The formula should go in cells K2:K5. Furthermore, after resolving the issue, amend the formula to consider blank cells as zeros, and format final output without decimals. ### spreadsheet_path /app/spreadsheets/1_38074_input.xlsx ### instruction_type Cell-Level Manipulation ### answer_position K2:K5 ### output_path /app/output/1_38074_output.xlsx You should generate Python code for the final solution of the question. ``` --- Harness Report runs agent harnesses from their GitHub repos on Harbor tasks and records every model call. Every page is also `.md` and `.json`; index: https://harnessreport.com/llms.txt · MCP: https://harnessreport.com/mcp