Model Monetary Data
Applications that handle monetary data often require the ability to capture fractional units of currency and need to emulate decimal rounding with exact precision when performing arithmetic. The binary-based floating-point arithmetic used by many modern systems (i.e., float, double) is unable to represent exact decimal fractions and requires some degree of approximation making it unsuitable for monetary arithmetic. This constraint is an important consideration when modeling monetary data.
There are several approaches to modeling monetary data in MongoDB using the numeric and non-numeric models.
The numeric model may be appropriate if you need to query the database for exact, mathematically valid matches or need to perform server-side arithmetic, e.g.,
$mul, and aggregation framework arithmetic.
The following approaches follow the numeric model:
If there is no need to perform server-side arithmetic on monetary data or if server-side approximations are sufficient, modeling monetary data using the non-numeric model may be suitable.
The following approach follows the non-numeric model:
- Using two fields for the monetary value: One field stores the exact monetary value as a non-numeric
stringand another field stores a binary-based floating-point (
doubleBSON type) approximation of the value.
New in version 3.4.
decimal BSON type uses the IEEE 754 decimal128 decimal-based floating-point numbering format. Unlike binary-based floating-point formats (i.e., the
double BSON type), decimal128 does not approximate decimal values and is able to provide the exact precision required for working with monetary data.
decimal values are assigned and queried using the
NumberDecimal() constructor. The following example adds a document containing gas prices to a
The following query matches the document above:
For more information on the
decimal type, see NumberDecimal.
A collection’s values can be transformed to the
decimal type by performing a one-time transformation or by modifying application logic to perform the transformation as it accesses records.
A collection can be transformed by iterating over all documents in the collection, converting the monetary value to the
decimal type, and writing the document back to the collection.
It is strongly advised to add the
decimal value to the document as a new field and remove the old field later once the new field’s values have been verified.
Be sure to test
decimal conversions in an isolated test environment. Once datafiles are created or modified with MongoDB version 3.4 they will no longer be compatible with previous versions and there is no support for downgrading datafiles containing decimals.
Scale Factor Transformation:
Consider the following collection which used the Scale Factor approach and saved the monetary value as a 64-bit integer representing the number of cents:
long value can be converted to an appropriately formatted
decimal value by multiplying
NumberDecimal("0.01") using the
$multiply operator. The following aggregation pipeline assigns the converted value to the new
priceDec field in the
The results of the aggregation pipeline can be verified using the
If you do not want to add a new field with the
decimal value, the original field can be overwritten. The following
update() method first checks that
price exists and that it is a
long, then transforms the
long value to
decimal and stores it in the
The results can be verified using the