Handling differences between MS-SQL and PostgreSQL money datatypes

Question

This is a follow-up/ to my previous question Is it possible to insert un-formatted data for money data type in PostgreSQL

1. I could read money data by typecasting it into numeric

    SELECT '52093.89'::money::numeric
   

   Will it give inconsistent data, rounded etc.. ?

2. As per ISO standard, the money datatype size is (19,4) but why PostgreSQL returns size as 2147483647? Is this size locale dependent?

3. Is it a good idea to use the money datatype to store/retrieve locale specific values? If not, when is the money datatype best suitable?

I have to use the money datatype for certain cases. My worry is will I get any problems if I treat money datatype values as decimal values?

I don't bother about locale specific values such as symbol, digit, and decimal grouping etc.. Can I store money datatype values as per lc_monetary settings and reading by typecasting it to numeric?

Answer 1

In Postgres, the cast from money to numeric is exact, immutable and does not round.

In SQL Server the data type MONEY uses 8 bytes to store a values in any currency in the numeric range from -922337203685477.5808 to 922337203685477.5807.
Matches numeric(19,4).

In PostgreSQL the data type money also occupies 8 bytes but the allowed range is -92233720368547758.08 to +92233720368547758.07. Same precision, but different scale by default (without LC_MONETARY setting other than 'C') - only two fractional digits.
Matches numeric(19,2).

But, quoting the manual:

The fractional precision is determined by the database's lc_monetary setting.

I updated the tag info for money accordingly.

As @a_horse commented (repeatedly), money is not defined by the SQL standard.

And it's pretty much consensus around here to stay away from the money data type if you can.