I agree, and I thought the same approach. However, my concern is more about %ValidateIndices() itself in InterSystems IRIS.

For large tables with heavily corrupted indexes, the bottleneck is not necessarily retaining tErrors after execution, but the fact that %ValidateIndices() builds large local arrays internally. If a single index contains a large number of corrupt entries, the process could still hit excessive memory usage or before control even returns to the caller.

The main point here is that the requirement is only to identify the corrupted index name, not to collect every corrupted entry into tErrors.

Currently there’s no lightweight mode to:

• stop after first/significant corruption
• or simply return whether an index is corrupted

without materializing all failed entries into local arrays.