Before I became an independent developer, I was engaged in statistical work on corporate financial data. Each month, the most time-consuming task was not the data statistics themselves but producing, in fixed formats, successive batches of reports for management to review. Management had high requirements for report layout, and in many cases reports needed to be printed on paper for convenient review. I had a certain understanding of the SQL query language, and the data statistics work was generally completed with ease; however, meeting management's requirements for reports consumed a substantial amount of time every month. Although fixed report templates existed, the statistical results had to be pasted manually, column by column, into Excel spreadsheets and Word documents. If the formatting was accidentally overwritten, it was necessary to adjust data formats, decimal places, merged headers, and content alignment, as well as report dates and other information—all relying on manual copy-paste and formatting adjustments. Producing one monthly report took an entire day, and any change in the calculation basis required redoing everything. I learned that almost all my colleagues were in the same situation: the data were available; what was lacking was a finished report that could be directly printed and submitted.
I investigated commercial software of the report designer type. Installation was heavy, learning costs were high, and templates were locked in proprietary formats that business users could not modify at will. I also tried mail merge and VBA form filling, which had very weak support for multiple datasets and dynamic rows. I therefore decided to develop a lightweight reporting tool by myself in accordance with my own usage habits. The appropriate approach was template plus marker: the template remained the Excel/Word file most familiar to users; placeholders such as {% marker %} were written in cells; and the program was responsible for parsing, querying data, backfilling, and preserving the original format.
The tool is designed around three layers of abstraction: data sources (MySQL/PostgreSQL/SQLite/SQL Server/Oracle/DuckDB), datasets (a SQL statement with parameters), and templates (Excel/Word/TXT with inserted markers). The marker syntax supports single-value references and entire data blocks, including horizontal and vertical filling, and includes a set of built-in inline functions for statistical scenarios. I will gradually compile the function list into documentation and will continue to expand it according to usage in due course. After users enter parameters and run the tool with one click, the output is a finished product that can be directly printed and submitted.
A typical use case for this tool is as follows: statistical staff write a statement such as SELECT month, metric FROM monthly_report and save it as a dataset; they insert several markers into their organization's Excel template; at runtime they enter the year; and within two minutes they obtain a fully completed monthly report for year-on-year analysis, ready for direct printing and use. Finance staff use it to merge multiple detail tables into a fixed report template; when the calculation basis changes, only the SQL statement needs to be modified, and not a single word in the template needs to be changed.
After the tool was completed and given to former colleagues for use, it basically solved the problems of repeated data retrieval and formatting in daily report preparation and saved considerable time.