Contact

FLOOR.PRECISE in Collabora Office: Round Toward the Lower Multiple

Math Intermediate Collabora Office Calc

Use Collabora FLOOR.PRECISE for signed lower bounds, interval bins and grouped-capacity examples with a fixed rounding direction.

Verification status: See the evidence and test scope described in this guide. How formula examples are checked

Quick answer: FLOOR.PRECISE rounds toward the lower multiple on the number line. =FLOOR.PRECISE(-17;5) is expected to return -20, and the sign of significance is ignored.

Availability and version history

  • Introduced: original Collabora release and launch date are not established by the sources reviewed.
  • Availability: documented in Collabora Office 26.04 help, scoped to the desktop-help release.
  • Platform scope: support in every online or mobile deployment is not inferred from desktop documentation.
  • Checked: October 14, 2026. Complete function-specific documentation and independent arithmetic controls; no native Collabora test was performed.

Syntax and fixed downward direction

=FLOOR.PRECISE(number[;significance])

The optional step defaults to 1. Positive inputs round toward zero; negative inputs round away from zero. There is no mode parameter.

The word “Precise” does not imply arbitrary-precision arithmetic. This is a named rounding rule with a sign-independent step, documented separately from FLOOR.

Examples use English names and semicolon separators. Step size and input units must describe the same quantity.

Example 1: lower multiples on both sides of zero

Enter -17, -15, -13, 13, 15 and 17 in A2:A7:

=FLOOR.PRECISE(A2;5)

Expected outputs are -20, -15, -15, 10, 15 and 15. Each result is no greater than the input. Exact multiples are unchanged.

Repeating the test with significance -5 gives the same outputs. The negative sign does not change the spacing’s magnitude or the direction of rounding.

Example 2: integer defaults

=FLOOR.PRECISE(7.4)
=FLOOR.PRECISE(-7.4)

Expected results are 7 and -8. The negative case distinguishes a mathematical floor from TRUNC, which would return -7 at its default precision.

If a negative toward-zero policy is required for a chosen multiple, use a deliberately selected FLOOR.MATH mode rather than adding an unsupported mode argument here.

Advanced example: label five-unit measurement bins

Enter measurements -0.5, 0.5, 4.9, 5 and 5.1 in D2:D6:

=FLOOR.PRECISE(D2;5)

Expected lower-bound labels are -5, 0, 0, 5 and 5. Under a lower-inclusive, upper-exclusive convention, these identify [-5,0), [0,5) and [5,10).

A boundary value of 5 belongs to the bin starting at 5. A value just below zero belongs to the bin starting at -5. The zero-based labels are numeric boundaries, not ordinal group numbers.

Advanced example: compare separate and pooled complete lots

Two locations each hold 9 units. A process accepts only lots of five:

=FLOOR.PRECISE(9;5)+FLOOR.PRECISE(9;5)
=FLOOR.PRECISE(9+9;5)

Expected usable totals are 10 and 15. Separate processing leaves eight units unused; pooling leaves three. Pooling is valid only when the real process permits combining stock.

Advanced check: bracket a negative measurement

For -17 with a five-unit grid, FLOOR.PRECISE returns -20, while CEILING.PRECISE returns -15. The original value lies between the two bounds.

Their difference is five for this nonmultiple, but zero for an exact multiple such as -15. Do not assume the two rounded bounds always differ by one full step.

Source scope and unchecked cases

The dedicated Collabora page explicitly defines the PRECISE syntax and signed direction. Its surrounding export paragraph discusses the unsuffixed FLOOR mapping, so this guide does not turn that paragraph into a separately certified FLOOR.PRECISE export result.

Zero significance, type conversion, native error labels and near-boundary floating-point results remain unverified. The examples establish the documented arithmetic rule using finite, nonzero steps and independently checked expectations.

References and verification

The complete function section was reviewed. Worked results are independently checked expectations; native execution and unlisted edge cases remain unverified.