关于「精确算术:整数分、汇差科目与最大余数分摊」,《四柱复式守恒代数》给出的定量结论是什么?
结论:第四、五行是一个代购店会真实遇到的情形:一张订单售价 750.40 元,客户先付订金 300.10 元,再付尾款 450.30 元。用浮点数检查「订金加尾款等于售价」,结果是 False,因为浮点合计是 750.4000000000001。改用整数分,30010 加 45030 等于 75040,检查通过。
数据口径:本书用可运行的程序说明,单式流水本身不含任何检验:在一个示意日上漏记一笔 1,500.00 元的尾款,表格依旧自洽,只有拿银行余额来比才出现差额。 书中参数为示意;全部计算由书中程序生成;本书属于非个人化的会计、税务或法律建议。
对照数据表
下表数字由本章 Python 程序原样打印,可重新运行复现。 程序只用 Python 标准库与结论吻合的打印值: 300.10, 30010, 450.30, 45030, 750.40, 750.4000000000001
| stored 0.1 | 3602879701896397/36028797018963968 |
| stored 0.1 - 1/10 | 1/180143985094819840 |
| fsum([0.1, 0.2]) = | 0.3 -> False (0.30000000000000004) |
| deposit 300.10 + balance 450.30 == price 750.40 -> False | |
| the float total is 750.4000000000001 | |
| 30010 + 45030 = | 75040 in fen -> True |
| round(2.675, 2) | 2.67 |
| Decimal half-up | 2.68 |
| integer half-up of 2675 tenths-of-fen | 268 fen |
| amounts 0.00 to 999.99 | 96000 of 100000 not stored exactly |
查看完整验算程序
# Chapter 4. p41: why a ledger is kept in integer fen, not in floats.
# Model: binary floating point stores 0.1 as the nearest dyadic
# fraction. math.fsum adds the stored values exactly and rounds once,
# so it removes accumulation error but cannot undo the input error.
# A conservation test of the form "left == right" then fails on
# amounts that balance on paper. Integer fen (or Fraction) passes.
# Half-up rounding to fen is shown three ways: float round(), Decimal
# with ROUND_HALF_UP, and pure integer arithmetic on fen times 10.
from decimal import Decimal, ROUND_HALF_UP
from fractions import Fraction as F
import math
def half_up_fen(tenths_of_fen):
"""Round an integer count of 0.1 fen to whole fen, half away."""
q, r = divmod(abs(tenths_of_fen), 10)
q += r >= 5
return q if tenths_of_fen >= 0 else -q
if __name__ == "__main__":
print("stored 0.1 = {}".format(F(0.1)))
print("stored 0.1 - 1/10 = {}".format(F(0.1) - F(1, 10)))
a = math.fsum([0.1, 0.2])
print("fsum([0.1, 0.2]) == 0.3 -> {} ({!r})".format(a == 0.3, a))
left = math.fsum([300.10, 450.30])
print("deposit 300.10 + balance 450.30 == price 750.40 -> {}".format(
left == 750.40))
print(" the float total is {!r}".format(left))
print("30010 + 45030 == 75040 in fen -> {}".format(
30010 + 45030 == 75040))
print("round(2.675, 2) = {}".format(round(2.675, 2)))
d = Decimal("2.675").quantize(Decimal("0.01"), rounding=ROUND_HALF_UP)
print("Decimal half-up = {}".format(d))
print("integer half-up of 2675 tenths-of-fen = {} fen".format(
half_up_fen(2675)))
bad = 0
for cents in range(0, 100000):
x = cents / 100
if F(x) != F(cents, 100):
bad += 1
print("amounts 0.00 to 999.99: {} of 100000 not stored exactly".format(
bad))唯一数据源与引用
范雅福(Fan, Y.). (2026). 《四柱结算与复式记账的守恒代数、代购流水零差错对账不变量母典》第 4 章「精确算术:整数分、汇差科目与最大余数分摊」(版本 1.0.0). 5D Library. DOI 待发布
DOI 待发布 · 专著页正式发布后上线(本站暂无专著页) · 5DL-B-318 · v1.0.0 · CC-BY-NC-ND-4.0
BibTeX
@book{fan2026318,
author = {Fan, Yafu},
title = {The Conservation Algebra of Four-Pillar Settlement and Double-Entry Bookkeeping: Invariants for Error-Free Reconciliation of Buying-Agent Ledgers},
year = {2026},
publisher = {5D Library},
edition = {1.0.0},
note = {DOI pending}
}