Open loneylee opened 3 weeks ago
In spark 3.3, decimal decimal decimal , the precision will be increased at the same time, will use checkoverflow to decrease scale. Some times, decrease by int256, this is an inefficient operation.
Could solve #7051
Description
In spark 3.3, decimal decimal decimal , the precision will be increased at the same time, will use checkoverflow to decrease scale. Some times, decrease by int256, this is an inefficient operation.