RAWSUBTRACT in Collabora Office: Inspect Subtraction Without Roundoff Suppression
Use Collabora RAWSUBTRACT in left-to-right order, reproduce exact controls, inspect tiny residuals, and distinguish binary rounding from decimal arithmetic.
Verification status: See the evidence and test scope described in this guide. How formula examples are checked
Quick answer: =RAWSUBTRACT(1000;125.5;44.25;5) has independently expected result 825.25. It subtracts each later argument from the first, in order, without eliminating small roundoff errors.
Syntax and evaluation order
=RAWSUBTRACT(Minuend; Subtrahend1 [; Subtrahend2] … [; Subtrahend254])
The Minuend is the starting number or a reference to a numeric cell. Each Subtrahend is another number or numeric cell reference to subtract. At least two arguments are required; the documented syntax allows up to 254 subtrahends after the initial value.
Collabora explicitly states that processing is left to right. For example, =RAWSUBTRACT(1;2;3;4) means ((1−2)−3)−4, with exact integer result −8. It does not mean 1−(2−3−4).
“Raw” does not mean infinite precision. The function exposes subtraction without suppressing small residuals; it does not make binary floating-point storage represent every decimal fraction exactly.
Example: reconcile a simple balance
Enter 1000 in A2, 125.5 in B2, 44.25 in C2, and 5 in D2. In E2 enter:
=RAWSUBTRACT(A2;B2;C2;D2)
The independent result is 825.25. The step-by-step balances are 874.5, 830.25, and 825.25. These quarter-unit values have straightforward exact binary representations, making this a clean arithmetic control.
Store each deduction in its own labeled cell. If B2 already contains a negative amount, subtracting B2 adds its magnitude. For instance, =RAWSUBTRACT(100;-20;5) has exact result 115. Agree on a sign convention before combining imported credits and debits.
Reproduce the vendor’s tiny-residual example
=RAWSUBTRACT(0.987654321098765;0.9876543210987)
The documented result is 6.53921361504217E−14. Display scientific notation or enough decimal places to see it. A cell formatted to two decimals can display zero even when the stored result is nonzero.
Exact decimal subtraction of the printed inputs is 6.5E−14. A separate binary64 arithmetic check gives approximately 6.539213615042172E−14, matching the vendor’s control at its shown precision. The difference illustrates decimal representation and binary arithmetic, rather than a reason to claim that the raw result is mathematically exact in decimal.
Advanced example: expose cancellation sensitivity
Consider these two formulas:
=RAWSUBTRACT(10000000000000000;10000000000000000;1)
=RAWSUBTRACT(10000000000000000;1;10000000000000000)
Their exact real-arithmetic result is −1 in both cases. A separate ordinary binary64 model gives −1 for the first ordering and 0 for the second: subtracting one from the very large initial number loses that small difference at the available precision.
Those are independent floating-point expectations, not captured Collabora results or a universal native guarantee. The example explains why the documented evaluation order matters and why reordering a calculation can affect numerical residuals.
For an audit, retain the original inputs and calculation order. A zero residual after cancellation is not automatically proof that the original values were exactly equal.
Advanced application: test a meaningful tolerance
If a reconciliation permits an absolute difference below 0.000001, calculate the residual once in E5 and test:
=ABS(E5)<0.000001
For the documented residual near 6.5392E−14, the independent test is TRUE. ABS removes the sign solely for this tolerance test. Choose the tolerance from the problem’s units and required precision, rather than adopting one number for every dataset.
ROUND can make a reporting value consistent with a chosen decimal precision, but it answers a different question from inspecting the unadjusted residual. Keep a raw residual and a rounded reporting column separate when both are needed.
Errors, limits, and compatibility
The explicit vendor error example supplies only one argument and returns Err:511, “Missing variable,” because at least two numbers are required. The section documents numbers and individual numeric cell references; it does not establish all range, text, blank, or other coercion behavior.
The function is identified as outside OpenDocument 1.3 OpenFormula, with namespace ORG.LIBREOFFICE.RAWSUBTRACT. That namespace is not evidence of identical native behavior in another product. Extreme values, underflow, and every possible error code are outside the checks in this guide.
Availability and verification
The exact function is documented in Collabora Office 26.04 desktop help, retrieved and reviewed on October 17, 2026. This establishes a documentation scope, not the first release that introduced the function or the last version in which it works; those versions and any retirement date remain unestablished. Collabora Online and mobile support are not inferred from desktop help, nor is behavior in another spreadsheet product.
The examples are documentation-based expectations with independent calculations. They have not been executed in Collabora Office. No application screenshot or native error/coercion certification is implied. Formulas use English function names and semicolon argument separators.
References
- Collabora Office 26.04: RAWSUBTRACT (checked October 17, 2026).
- Editorial testing standards.