How to convert Python decimal to SQLite numeric?

问题:

I have a program that reads financial data in JSON and inserts it into an SQLite database. The problem is when I'm inserting it into SQLite numeric column and it doesn't seem to like the decimal object.

I've found this question answered before, but the answer is outdated and from what I understand SQLite now has a currency data type called numeric.

Right now as a workaround I'm storing decimal values as text, but is it possible to store it as numeric? Am I stuck with the overhead of converting decimals to strings and vice versa for database inserts and financial calculations?

回答1:

sqlite3 allows you to register an adapter (to transparently convert Decimals to TEXT when inserting) and a converter (to transparently convert TEXT into Decimals when fetching).

The following is a lightly modified version of the example code from the docs:

import sqlite3
import decimal
D=decimal.Decimal

def adapt_decimal(d):
    return str(d)

def convert_decimal(s):
    return D(s)

# Register the adapter
sqlite3.register_adapter(D, adapt_decimal)

# Register the converter
sqlite3.register_converter("decimal", convert_decimal)

d = D('4.12')

con = sqlite3.connect(":memory:", detect_types=sqlite3.PARSE_DECLTYPES)
cur = con.cursor()
cur.execute("create table test(d decimal)")

cur.execute("insert into test(d) values (?)", (d,))
cur.execute("select d from test")
data=cur.fetchone()[0]
print(data)
print(type(data))

cur.close()
con.close()

yields

4.12
<class 'decimal.Decimal'>

回答2:

At least the page you linked to didn't mention a currency data type, and the decimal(10,5) example datatype simply turns on the NUMERIC affinity.

If it were my data, I'd probably store an integer number of currency "units" -- pennies, or tenths of a penny, or hundredths of a penny, whatever is appropriate -- and then use a scaling factor to turn the integer input into the decimal equivalent to compute with when reading data from the database. It's harder to screw up with integers.

回答3:

I found that I had to make a small tweak to unutbu's approach. With a modified example with a value '4.00', it comes back out of the database as '4'. I'm dealing with commodities and don't want to hardcode the precision into the database (like I'd be doing if I just multiplied and divided by 100). So I tweaked the conversion functions as follows:

def adapt_decimal(d):
    return '#'+str(d)

def convert_decimal(s):
    return D(s[1:])

which isn't aesthetically great but does defeat sqlite's eagerness to store the field as an integer and lose track of the precision.

回答4:

http://www.sqlite.org/datatype3.html#affinity

• If the declared type of the column contains any of the strings "CHAR", "CLOB", or "TEXT" then that column has TEXT affinity. Notice that the type VARCHAR contains the string "CHAR" and is thus assigned TEXT affinity.

You can declare columns to be any type you like:

CREATE TABLE a_test( a_decimal DECTEXT NOT NULL /* will be stored as TEXT */ );

---

def adapt_decimal(d):
    return str(d)

def convert_decimal(s):
    return decimal.Decimal(s)

# Register the adapter
sqlite3.register_adapter(decimal.Decimal, adapt_decimal)

# Register the converter
sqlite3.register_converter("DECTEXT", convert_decimal)

con = sqlite3.connect("test.s3db", detect_types=sqlite3.PARSE_DECLTYPES)
C1 = con.cursor()

C1.execute("INSERT INTO a_test VALUES(?)", (decimal.Decimal("102.20"),))

Don't know if this is a good way to handle it or not - comments welcome