3693075 - Managing Memory Growth in Technical Tables via Inverted Individual Indexes

In SAP HANA systems, high memory consumption in certain technical tables can be observed, which is not caused by data volume but by the use of inverted value indexes. 

These index types commonly used on tables with multi-column primary keys can cause significant memory overhead, as they store a concatenated attribute for indexing purposes.

Even though regular housekeeping and data cleanup jobs (e.g., E2E BI housekeeping) were running successfully, memory usage remained high due to the inefficient structure of these inverted value indexes.

To analyze and confirm the issue, the following SQL scripts (SAP Note 1969700) were used:

• HANA_Memory_Overview – Identified tables with high overall memory consumption.
• HANA_Tables_Largest_Tables – Showed that a large portion of memory was consumed by indexes rather than table data.
• HANA_Indexes_Overview – Revealed that affected tables were using inverted value indexes (INDEX_TYPE) and showed the potential memory savings (MEM_CONC_GB) if converted to inverted individual indexes.

Additionally, running the SAP HANA Mini Checks can help highlight potential index-related memory issues. In particular, Check ID M0452 (“Size of unique indexes (%)”) can be used to identify tables where unique indexes occupy a large percentage of the total memory footprint, further pointing to optimization opportunities.

SAP HANA, platform edition 2.0 and above

 Inverted Individual Indexes, memory consumption, index size, data cleanup , KBA , HAN-DB , SAP HANA Database , Problem