expression: extend locale support for format()

[dveeden]

What problem does this PR solve?

Issue Number: close #56167

Problem Summary:

What changed and how does it work?

Check List

Tests

• Unit test
• Integration test
• Manual test (add detailed scripts or steps below)
• No need to test

  • I checked and no code files have been changed.

Side effects

• Performance regression: Consumes more CPU
• Performance regression: Consumes more Memory
• Breaking backward compatibility

Documentation

• Affects user behaviors
• Contains syntax changes
• Contains variable changes
• Contains experimental features
• Changes MySQL compatibility

Release note

Please refer to Release Notes Language Style Guide to write a quality release note.

The FORMAT() function now supports more locales

[ti-chi-bot]

Skipping CI for Draft Pull Request.
If you want CI signal for your change, please convert it to an actual PR.
You can still manually trigger a test run with /test all

[tiprow]

Hi @dveeden. Thanks for your PR.

PRs from untrusted users cannot be marked as trusted with /ok-to-test in this repo meaning untrusted PR authors can never trigger tests themselves. Collaborators can still trigger tests on the PR using /test all.

I understand the commands that are listed here.

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository.

[dveeden]

/ok-to-test

[codecov]

Codecov Report

Attention: Patch coverage is 58.82353% with 21 lines in your changes missing coverage. Please review.

Project coverage is 75.4794%. Comparing base (f399e91) to head (e4c693e).
Report is 354 commits behind head on master.

Additional details and impacted files

@@               Coverage Diff                @@
##             master     #56168        +/-   ##
================================================
+ Coverage   73.4247%   75.4794%   +2.0546%     
================================================
  Files          1619       1628         +9     
  Lines        446765     463235     +16470     
================================================
+ Hits         328036     349647     +21611     
+ Misses        98618      92892      -5726     
- Partials      20111      20696       +585     

Flag	Coverage Δ
integration	48.8034% <56.8627%> (?)

Flags with carried forward coverage won't be shown. Click here to find out more.

Components	Coverage Δ
dumpling	52.9478% <ø> (ø)
parser	∅ <ø> (∅)
br	52.2601% <ø> (+6.7308%)	⬆️

[kennytm]

👀 i'm afraid the test cases is too simple, since format() is supposed to handle thousands separators as well so you should at least test

mysql> select format('12345678.9', 3, 'en_US');
12,345,678.900
mysql> select format('12345678.9', 3, 'nl_NL');
12345678,900
mysql> select format('12345678.9', 3, 'id_ID');
12.345.678,900
mysql> select format('12345678.9', 3, 'it_CH'); -- Switzerland
12'345'678,900
mysql> select format('12345678.9', 3, 'de_CH'); -- Switzerland
12'345'678.900
mysql> select format('12345678.9', 3, 'en_IN'); -- India
1,23,45,678.900

[lcwangchao]

Why ignore error here? I think it's better to handle error even if we suppose it should not happen

[ti-chi-bot]

[APPROVALNOTIFIER] This PR is NOT APPROVED

This pull-request has been approved by:
Once this PR has been reviewed and has the lgtm label, please assign d3hunter, xuhuaiyu for approval, ensuring that each of them provides their approval before proceeding. For more information see the Code Review Process.

The full list of commands accepted by this bot can be found here.

Needs approval from an approver in each of these files:

• OWNERS
• pkg/expression/OWNERS
• pkg/parser/OWNERS

Approvers can indicate their approval by writing /approve in a comment
Approvers can cancel approval by writing /approve cancel in a comment

[dveeden]

/test all

[kennytm]

why will it end with ....372529029846191406250? the input should be a decimal number with exact decimal representation 🤔

mysql> select FORMAT(12345678.9, 99999999999999999, 'de_CH');
+------------------------------------------------+
| FORMAT(12345678.9, 99999999999999999, 'de_CH') |
+------------------------------------------------+
| 12'345'678.900000000000000000000000000000      |
+------------------------------------------------+
1 row in set (0.00 sec)

also why is it separated using ’ U+2019 RIGHT SINGLE QUOTATION MARK rather than ' U+0027 APOSTROPHE? is this OS-dependent?

[dveeden]

Go

package main

import (
	"golang.org/x/text/language"
	"golang.org/x/text/message"
	"golang.org/x/text/number"
)

func main() {
	p := message.NewPrinter(language.MustParse("de_CH"))
	p.Printf("%f", number.Decimal(12345678.9, number.Scale(1)))
}

https://go.dev/play/p/7Iv3SA0R-bR

output:

12’345’678.9

C with ICU

#include <stdio.h>
#include <unicode/unum.h>
#include <unicode/uloc.h>
#include <unicode/ustring.h>
#include <unicode/ustdio.h>

int main() {
    double number = 12345678.9;
    const char* locale = "de_CH";
    UChar result[100];

    UErrorCode status = U_ZERO_ERROR;
    UNumberFormat* fmt = unum_open(UNUM_DECIMAL, NULL, 0, locale, NULL, &status);
    int32_t resultLength = sizeof(result) / sizeof(result[0]);
    resultLength = unum_formatDouble(fmt, number, result, resultLength, NULL, &status);
    char resultUTF8[100];
    u_strToUTF8(resultUTF8, sizeof(resultUTF8), NULL, result, resultLength, &status);
    printf("%s\n", resultUTF8);
    unum_close(fmt);
    return 0;
}

output:

12’345’678.9

Python

#!/bin/python3
import locale

locale.setlocale(locale.LC_NUMERIC, "de_CH")
print(locale.format_string("%.2f", 12345678.9, grouping=True))

output:

12'345'678.90

Conclusion

Depending on the OS, programming language and libraries either ’ or ' are used. Looks like both are acceptable.

Note that this results in minor differences between TiDB and MySQL (tested with 9.0.1):

locale	TiDB	MySQL	Difference?
nl_BE	12.345.678,9000	12345678,9000	Yes
fr_BE	12 345 678,9000	12345678,9000	Yes
en_IN	1,23,45,678.9000	1,23,45,678.9000	No
de_CH	12’345’678.9000	12'345'678.9000	Yes
no_NO	12,345,678.9000	12.345.678,9000	Yes (big difference)
bg_BG	12 345 678,9000	12 345 678,9000	No
ar_SA	12,345,678.9000	12345678.9000	Yes
es_MX	12,345,678.9000	12,345,678.9000	No

[dveeden]

The no_NO difference is probably not a real issue as it should have been nb_NO or nn_NO instead of no_NO. See also https://bugzilla.redhat.com/show_bug.cgi?id=532487#c0

mysql-9.0.1> \W
Show warnings enabled.
mysql-9.0.1> SELECT FORMAT(12345678.9,4,'nb_NO');
+------------------------------+
| FORMAT(12345678.9,4,'nb_NO') |
+------------------------------+
| 12.345.678,9000              |
+------------------------------+
1 row in set (0.00 sec)

mysql-9.0.1> SELECT FORMAT(12345678.9,4,'nn_NO');
+------------------------------+
| FORMAT(12345678.9,4,'nn_NO') |
+------------------------------+
| 12,345,678.9000              |
+------------------------------+
1 row in set, 1 warning (0.00 sec)

Warning (Code 1649): Unknown locale: 'nn_NO'

mysql-8.0.11-TiDB-v8.4.0-alpha-295-g5f8ee5d509> \W
Show warnings enabled.
mysql-8.0.11-TiDB-v8.4.0-alpha-295-g5f8ee5d509> SELECT FORMAT(12345678.9,4,'nb_NO');
+------------------------------+
| FORMAT(12345678.9,4,'nb_NO') |
+------------------------------+
| 12 345 678,9000              |
+------------------------------+
1 row in set (0.01 sec)

mysql-8.0.11-TiDB-v8.4.0-alpha-295-g5f8ee5d509> SELECT FORMAT(12345678.9,4,'nn_NO');
+------------------------------+
| FORMAT(12345678.9,4,'nn_NO') |
+------------------------------+
| 12 345 678,9000              |
+------------------------------+
1 row in set (0.01 sec)

[dveeden]

/test all

[tiprow]

@dveeden: The following test failed, say /retest to rerun all failed tests or /retest-required to rerun all mandatory failed tests:

Test name	Commit	Details	Required	Rerun command
fast_test_tiprow	e4c693e	link	true	/test fast_test_tiprow

Full PR test history. Your PR dashboard.

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes-sigs/prow repository. I understand the commands that are listed here.

[ti-chi-bot]

@dveeden: The following tests failed, say /retest to rerun all failed tests or /retest-required to rerun all mandatory failed tests:

Test name	Commit	Details	Required	Rerun command
idc-jenkins-ci-tidb/mysql-test	e4c693e	link	true	/test mysql-test
idc-jenkins-ci-tidb/unit-test	e4c693e	link	true	/test unit-test

Full PR test history. Your PR dashboard.

Instructions for interacting with me using PR comments are available here. If you have questions or suggestions related to my behavior, please file an issue against the kubernetes/test-infra repository. I understand the commands that are listed here.

[dveeden]

I'm going to part this pull request for now. The reason for this is that golang.org/x/text can do nice formatting for a large number of locales, but it only accepts integer and float types, which cause incorrect results due to how floats are handled.

I've filed a proposal to allow direct conversion from a string instead of relying on first converting the value to a float:
golang/go#69740

The options we have now are:

1. Wait for golang.org/x/text improvements
2. Do the locale formatting without an external library. This is quite a bit of work to do it correctly. However the benefit is that we more easily follow the output of MySQL more closely.
3. Find another library that does this

cc @kennytm @lcwangchao @YangKeao