Just to be clear, I was not recommending a change, just offering an alternative.ismangil wrote:Maybe it will be more robust, but in this particular issue isn't the case that the transaction datetime happens well within the floor on the history query?
If you know for a fact the transaction begin date stored in the database occurs *after* the history's floor date (which in this case is based on the same system clock which created the transaction's begin date), then it follows the data should be retrieved.