1. Overview (Project Background)

This project was initiated to convert SQL used in BI tools such as Redash and Tableau to suit a new environment.
The existing SQL had complex structures and mixed various functions, making automatic conversion alone insufficient. Additionally, the Presto execution environment was operated based on VPN access, causing continuous communication and network delay issues.

In particular, there was a requirement to convert approximately 4,000 SQL queries within 8 weeks under a limited timeline, making it essential to establish an efficient conversion strategy and work methodology.

2. Solution (Resolution Approach)

To improve the efficiency of large-scale SQL conversion work, a Python-based automation program was developed, and repetitive tasks were minimized around this program.
Additionally, a close collaboration system was established with the client's personnel to quickly share query characteristics and business importance.

Clear priorities were defined for the SQL conversion targets, and work was conducted in stages starting from core dashboards and key analysis queries, enabling maximum conversion within the limited timeframe.

3. Result (Achievements)

As a result of the conversion work, some of the overall targets remained in a state requiring additional review and refinement.
A total of 648 queries (16.2%) were identified as requiring continuous conversion work due to complexity and structural limitations, and 521 queries (13.0%) were classified as excluded or lower priority considering business importance and utilization.

Additionally, the executability of the converted SQL was verified, and targets requiring additional work for future performance improvement and optimization were also identified. Through this process, beyond simple conversion, follow-up tasks for improving SQL quality and operational efficiency were clearly organized.