also, check for scale >= 0 && scale <= precision ?

I'm not sure. Why can't scale be arbitrary? It's simply an exponent.

From this

Precision is the number of digits in a number. Scale is the number of digits to the right of the decimal point in a number. For example, the number 123.45 has a precision of 5 and a scale of 2.

The number of digits after the decimal must be >=0 and must be <= the digits in the number. Am I missing something ?

This is Microsoft-specific. There is no a priori reason why scale should be limited. Apparently Oracle allows scales between -128 and 127 (?):
https://docs.oracle.com/cd/A81042_01/DOC/server.816/a76965/c10datyp.htm#743

the 0 to 38 thing is pretty ubiquitous in the database landscape

• Spark SQL https://spark.apache.org/docs/2.0.2/api/java/org/apache/spark/sql/types/DecimalType.html
• Impala https://github.com/apache/impala/blob/master/be/src/udf/udf.h#L664
• Hive https://cwiki.apache.org/confluence/display/Hive/LanguageManual+Types#LanguageManualTypes-DecimalsdecimalDecimals
• Presto https://prestodb.github.io/docs/current/functions/decimal.html

MapD seems to be capped at 19 digits of precision instead of 38, presumably to fit in 64 bits https://www.omnisci.com/docs/latest/5_datatypes.html

@pitrou - thanks, I didn't realize that oracle allows scale to be more than precision, and -ve also. The decimal functions in gandiva don't handle this.

However, I checked the links from @wesm - spark-sql, impala and presto (and of course, sql-server) all require scale to be <= precision.