This updates some of the sub unit values #12

matthewsnell · 2019-10-25T11:25:48Z

This fixes the issue #9
Values were taken from https://developers.facebook.com/docs/marketing-api/currencies/

nickyr

Hey! I really appreciate you taking the time to do this.

Of course, because currencies are a pain, Facebook's list and the list @shridharama provided in the original ticket don't agree on all currencies. I commented where I saw discrepancies.

Any thoughts from either of you about how we should address this? The ISO link may be more technically accurate, but the Facebook link may be more useful in practice. I'm sort of leaning towards using the mentality "Facebook knows more than me" and accepting all the proposed changes here, but I'd love input from others.

nickyr · 2019-10-25T13:50:19Z

money/currency.py

@@ -401,7 +401,7 @@ class CurrencyHelper:
            'display_name': 'Colombian Peso',
            'numeric_code': 170,
            'default_fraction_digits': 2,
-            'sub_unit': 100,


Facebook might be wrong on this one. The link @shridharama provided and Wikipedia both indicate that 100 is correct.

It's possible Facebook isn't "wrong" here per se but is just setting arbitrary rules for the minimum allowed bid on their ad units, which sometimes matches up with established convention for that currency and sometimes doesn't (see here in the docs). In that case, if the sub unit isn't defunct we may want to go with the conventional usage as opposed to Facebook's rules. Thoughts?

Ooh good find, @kaylafuchs. @matthewsnell, maybe we should leave Colombia and Costa Rica as they were and take your changes on the others?

nickyr · 2019-10-25T13:51:14Z

money/currency.py

@@ -413,7 +413,7 @@ class CurrencyHelper:
            'display_name': 'Costa Rican Colon',
            'numeric_code': 188,
            'default_fraction_digits': 2,
-            'sub_unit': 100,


nickyr · 2019-10-25T13:52:50Z

money/currency.py

@@ -575,13 +575,13 @@ class CurrencyHelper:
            'display_name': 'Forint',
            'numeric_code': 348,
            'default_fraction_digits': 2,
-            'sub_unit': 100,


now for this one, the two lists disagree but Wikipedia says the sub unit is defunct, so maybe Facebook's is more accurate here

Was in Hungary this summer & did not ever encounter any denomination smaller than a Forint. Inflation is crazy there so it makes sense that the sub unit would be defunct. Based on those Facebook docs it looks like 1 indicates that there is no sub unit allowed, so it seems they actually agree here!

nickyr · 2019-10-25T13:53:40Z

money/currency.py

        },
        Currency.IDR: {
            'display_name': 'Rupiah',
            'numeric_code': 360,
            'default_fraction_digits': 2,
-            'sub_unit': 100,


here's another where wikipedia says the subunit is obsolete

nickyr · 2019-10-25T13:55:09Z

money/currency.py

@@ -1073,7 +1073,7 @@ class CurrencyHelper:
            'display_name': 'New Taiwan Dollar',
            'numeric_code': 901,
            'default_fraction_digits': 2,
-            'sub_unit': 100,


For this one, the lists disagree and wikipedia says the subunit is "rarely" used

nickyr · 2019-10-25T13:59:50Z

I'll definitely release this with a new major version

matthewsnell · 2019-10-26T18:23:20Z

I would probably lean more towards using Facebook's values as well.

nickyr · 2019-10-28T13:54:08Z

@shridharama @jcmanzo @kaylafuchs @DanayaMel @brendanpeters @lorenzocastillo I will merge this on Wednesday if nobody objects.

kaylafuchs · 2019-10-28T14:49:35Z

No objections here! I would be inclined to trust Facebook API docs over Wikipedia but it would be nice to have a more authoritative source for this kind of thing. 🤔

Edit: after looking more closely at those Facebook docs it seems that they're tweaking some currency rules in order to set the minimum allowed bid on their ad units. In that case we may not want to trust them.

jcmanzo · 2020-01-29T18:30:37Z

Looks like this never got merged. I'll take care of it later today..

kaylafuchs · 2020-01-29T18:35:56Z

@jcmanzo I think there's still an open question here about trusting the Facebook docs for Colombia and Costa Rica because they're using sub units to set a minimum allowed bid on their ad units - https://developers.facebook.com/docs/marketing-api/currencies/#example-offset-1

suspectpart · 2020-10-18T08:23:59Z

I guess the canonical source of truth here should be ISO 4217 which delineates currency designators, country codes and sub units. Wikipedia states:

ISO 4217 is a standard first published by International Organization for Standardization in 1978, which delineates currency designators, country codes (alpha and numeric), and references to minor units in three tables [...]. The ISO 4217 code list is used in banking and business globally. In many countries the ISO codes for the more common currencies are so well known publicly that exchange rates published in newspapers or posted in banks use only these to delineate the currencies, instead of translated currency names or ambiguous currency symbols. ISO 4217 codes are used on airline tickets and international train tickets to remove any ambiguity about the price.

Even when you encounter sub units rarely in some countries (e.g. for the Forint in Hungary), we should still stick to the Standard if it nominally allows sub units.

The latest version of the Standard was published on 29 August, 2018 and can be found here as an XML or XLS file.

I ran a small script that loads the XLS file, runs every currency through the CurrencyHelper and compares the sub units to the ones specified in the CurrencyHelper. These are the results:

ISO_4217 (Published 2018-08-29)
----------------
Can't process ANTARCTICA (No universal currency)
Can't process PALESTINE, STATE OF (No universal currency)
Can't process SOUTH GEORGIA AND THE SOUTH SANDWICH ISLANDS (No universal currency)
[BIF] expected 1, got 100
[CLF] expected 10000, got 100
[CLP] expected 1, got 100
[DJF] expected 1, got 100
[GNF] expected 1, got 100
[ISK] expected 1, got 100
[KMF] expected 1, got 100
[KRW] expected 1, got 100
[MGA] expected 100, got 5
[MRU] Currency Missing! (expected 100)
[PYG] expected 1, got 100
[RWF] expected 1, got 100
[STN] Currency Missing! (expected 100)
[UGX] expected 1, got 100
[UYI] expected 1, got 100
[UYW] Currency Missing! (expected 10000)
[VES] Currency Missing! (expected 100)
[VND] expected 1, got 10
[XAF] expected 1, got 100
[XAG] expected 'N.A.', got 100
[XAU] expected 'N.A.', got 100
[XBA] expected 'N.A.', got 100
[XBB] expected 'N.A.', got 100
[XBC] expected 'N.A.', got 100
[XBD] expected 'N.A.', got 100
[XDR] expected 'N.A.', got 100
[XOF] expected 1, got 100
[XPD] expected 'N.A.', got 100
[XPF] expected 1, got 100
[XPT] expected 'N.A.', got 100
[XSU] expected 'N.A.', got 100
[XTS] expected 'N.A.', got 100
[XUA] expected 'N.A.', got 100
[XXX] Currency Missing! (expected N.A.)

I would suggest to approach this as follows:

Discuss whether whether 100 is the correct sub units value for all X currencies (those are not real national currencies, but rather a mixture of supranational currencies and procedural purposes)
Provide the missing currencies

[MRU] Currency Missing! (expected 100)
[STN] Currency Missing! (expected 100)
[UYW] Currency Missing! (expected 10000)
[VES] Currency Missing! (expected 100)
Be careful about MRU (Mauritanian ouguiya): This currency succeeds MRO and is one of two circulating currencies, along with the Malagasy ariary (MGA), whose division units are not based on a power of ten. The sub units here must also be 5. Note that in the ISO 4217 file, MRU and MGA are stated as sub units 2 (=100) but with a remark that it actually must be 1/5.)

Fix all currencies where the sub units differ from the ones stated in ISO 4217. That would be:

[BIF] expected 1, got 100
[CLF] expected 10000, got 100
[CLP] expected 1, got 100
[DJF] expected 1, got 100
[GNF] expected 1, got 100
[ISK] expected 1, got 100
[KMF] expected 1, got 100
[KRW] expected 1, got 100
[PYG] expected 1, got 100
[RWF] expected 1, got 100
[UGX] expected 1, got 100
[UYI] expected 1, got 100
[VND] expected 1, got 10

If we implement these changes, we could add a remark to the README.md stating that py-money complies with ISO 4217, making it suitable for finance and business applications that need to comply with this Standard.

We could create an integration test that runs the ISO 4217 file through the CurrencyHelper to assert compliance. If it is ever going to change in the future, you just drop the new specification and get instant feedback about necessary adjustments.

matthewsnell added 2 commits October 25, 2019 12:20

Update sub units

9593ddf

Update sub units

6101631

matthewsnell marked this pull request as ready for review October 25, 2019 11:26

nickyr reviewed Oct 25, 2019

View reviewed changes

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

This updates some of the sub unit values #12

This updates some of the sub unit values #12

matthewsnell commented Oct 25, 2019 •

edited

Loading

nickyr left a comment

nickyr Oct 25, 2019

kaylafuchs Oct 28, 2019

nickyr Oct 28, 2019

nickyr Oct 25, 2019

nickyr Oct 25, 2019

kaylafuchs Oct 28, 2019

nickyr Oct 25, 2019

nickyr Oct 25, 2019

nickyr commented Oct 25, 2019

matthewsnell commented Oct 26, 2019

nickyr commented Oct 28, 2019

kaylafuchs commented Oct 28, 2019 •

edited

Loading

jcmanzo commented Jan 29, 2020

kaylafuchs commented Jan 29, 2020

suspectpart commented Oct 18, 2020 •

edited

Loading

This updates some of the sub unit values #12

Are you sure you want to change the base?

This updates some of the sub unit values #12

Conversation

matthewsnell commented Oct 25, 2019 • edited Loading

nickyr left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

nickyr commented Oct 25, 2019

matthewsnell commented Oct 26, 2019

nickyr commented Oct 28, 2019

kaylafuchs commented Oct 28, 2019 • edited Loading

jcmanzo commented Jan 29, 2020

kaylafuchs commented Jan 29, 2020

suspectpart commented Oct 18, 2020 • edited Loading

matthewsnell commented Oct 25, 2019 •

edited

Loading

kaylafuchs commented Oct 28, 2019 •

edited

Loading

suspectpart commented Oct 18, 2020 •

edited

Loading