I have a database that has multiple tables without any clustered index. All that I have found are small (< 500 rows). Specifically, PRIMARY KEY NONCLUSTERED
with no other indexes at all.
At a very high level I understand the performance limitation of not having a clustered index. Perhaps more so I understand it isn't best practice, but before I aggressively start converting non-clustered primary keys to clustered ones, are there any caveats I should be aware of when converting these indexes?
The only one I can think of would be the time it would take to update an index on a large table, but again these are relatively small. Perhaps files & partitions, but I only have a single partition.
For reference, this is a 2008-compatible SQL Server database running on SQL Server 2016. This is in our primary OLTP system (these table are not involved in warehousing & ETL).