From Ten Minutes to Under One: A Financial Reporting Story

In a financial system I worked on with the WideSpace team, important reports took three to ten minutes, and some struggled as data grew. The problem was not limited to the Export PDF button; it involved the path from the database to the final document.

Start with the queries

We moved aggregation and filtering into stored procedures and temporary tables with indexes on relevant fields. This reduced the work passed back to the application. Query times improved substantially for the specific reports I discussed publicly.

Then handle generation

We used Snappy in place of DomPDF for long reports and Arabic text. Very large documents were built in sections and assembled with FPDI and PDF merging tools, helping with memory usage and usability.

In the cases I reported, generation went from around ten minutes to under one minute. That figure belongs to those reports and their test environment. The lasting lesson for me was to measure the whole workflow before changing a single library or adding more server capacity.