This is a bug.
Here's why this happens:
The error condition described in this method is hit:
https://docs.python.org/3/library/decimal.html#decimal.Decimal.quantize
"Unlike other operations, if the length of the coefficient after the
quantize operation would be greater than precision, then an
InvalidOperation is signaled. This guarantees that, unless there is an
error condition, the quantized exponent is always equal to that of the
right-hand operand."
What's happening is that the precision automatically inferred from the
numbers you've provided (the ones with the large fraction) exceed that of
the context.
e.g. with this patch it runs:
bergamot [git|v2]:~/p/beancount$ git diff
diff --git a/beancount/core/display_context.py
b/beancount/core/display_context.py
index cfed0188..9a9dbfce 100644
--- a/beancount/core/display_context.py
+++ b/beancount/core/display_context.py
@@ -217,6 +217,10 @@ class DisplayContext:
# Note: We could probably logging.warn() this situation here.
return number
qdigit = Decimal(1).scaleb(-num_fractional_digits)
+
+ import decimal; decimal.getcontext().prec = 30
+ print("XXX", qdigit, decimal.getcontext().prec)
+
return number.quantize(qdigit)
I'll have to find a solution.
On Sun, Nov 15, 2020 at 12:26 PM François <[email protected]> wrote:
> Hello,
>
> There is some strange behaviour with include directive.
>
> For example with those two files
>
> cat <<EOF >example.beancount
> 1970-01-01 open Expenses:MyExpense
> 1970-01-01 open Assets:MyAccount
>
> 1970-01-02 *
> Expenses:MyExpense 1000.00 EUR
> Assets:MyAccount -1000.00 EUR
>
> 1970-01-02 * "Verif "
> Expenses:MyExpense 333.0000000000000182076576039 EUR
> Assets:MyAccount -333.0000000000000182076576039 EUR
> EOF
>
> cat <<EOF >example_with_include.beancount
> include "example.beancount"
> EOF
>
> The results differ when we use the include file or not.
>
> In the case without include (bean-query example.beancount 'SELECT *')
> I have a stack trace with this error
> decimal.InvalidOperation: [<class 'decimal.InvalidOperation'>]
>
>
> In the case with include I have an output
>
> $ bean-query example_with_include.beancount 'SELECT *'
> date f narrat position
> ---------- - - ------ -----------------------------------
> 1970-01-02 * 1000.00 EUR
> 1970-01-02 * -1000.00 EUR
> 1970-01-02 * Verif 333.0000000000000182076576039 EUR
> 1970-01-02 * Verif -333.0000000000000182076576039 EUR
>
>
>
>
>
>
> I stumbled on this while splitting some beancount ledger on multiple
> files. In my case the transaction with a lot of decimal is generated by
> a plugin which makes some division of existing postings to split them
> between several accounts.
> dest_amount=core.amount.mul(original_posting.units,
> core.number.D(1/3))
>
> In this case I have the following outputs :
>
> $ bean-query exPlugin.beancount 'SELECT * WHERE account ~ ":A"'
> date f narrat position
> ---------- - - ------ ----------
> 1970-01-02 * Verif 333.33 EUR
>
> $ bean-query exPlugin_with_include.beancount 'SELECT * WHERE account ~
> ":A"'
> date f narrat position
> ---------- - - ------ ---------------------------------
> 1970-01-02 * Verif 333.3333333333333148296162562 EUR
>
>
> The directive are exactly the same in both cases. The only change is
> that in the first case I have only one file and in the second case I
> have the transaction in an included file.
>
> I like more the first version (with proper rounding) but what triggers
> me is the lack of coherence.
>
>
> Have you any explanation or workaround about this ?
>
>
> I use Beancount 2.2.3+hg20200223.0.3af921a
>
> François
>
> --
> You received this message because you are subscribed to the Google Groups
> "Beancount" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to [email protected].
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/beancount/20201115172254.7xtksqidbiwp2kjd%40fjo-extia-HPdeb.example.avalenn.eu
> .
>
--
You received this message because you are subscribed to the Google Groups
"Beancount" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
To view this discussion on the web visit
https://groups.google.com/d/msgid/beancount/CAK21%2BhPdXm5V1A9qk%3DthqjfWCk4%3DRD3jPs4DdJ3zF0asH%2Bo6XA%40mail.gmail.com.
