Calculation results in SQL Server retain decimals

Always keep decimals, calculation in the program is too cumbersome, but also to convert the operation. It is convenient to operate directly in the database.
The data type is defined as Decimal/numeric type, the decimal place needs to be arbitrarily set, divisor and dividend must have a type in decimal/numeric,
The following example:
SELECT 24.0000/38.0000--0
SELECT CAST (24/38 as NUMERIC (18,4))--0
SELECT CAST (24/38.0000 as NUMERIC (18,4))--0.6316 This result has been rounded
SELECT 24.0000/38.0000--0.6316

Decimal/numeric and int type operations result in the Decimal/numeric type because the decimal/numeric has a higher precedence than int.

Refer to the Type Priority table on MSDN:

SQL Server 2005 uses the following order of precedence for data types:

2. User-defined data type (highest)
   
   
4. sql_variant
   
   
6. XML
   
   
8. datetime
   
   
10. smalldatetime
    
    
12. float
    
    
14. Real
    
    
16. decimal
    
    
18. Money
    
    
20. smallmoney
    
    
22. bigint
    
    
24. int
    
    
26. smallint
    
    
28. tinyint
    
    
30. bit
    
    
32. ntext
    
    
34. text
    
    
36. Image
    
    
38. timestamp
    
    
40. uniqueidentifier
    
    
42. nvarchar(including nvarchar (max))
    
    
44. nchar
    
    
46. varchar (including varchar (max))
    
    
48. Char
    
    
50. varbinary(including varbinary (max))
    
    
52. binary(lowest)
    

Calculation results in SQL Server retain decimals