Contact

FLOOR.PRECISE in Google Sheets: Choose the Mathematical Lower Grid Point

Math Intermediate Google Sheets

Round down consistently in Sheets with FLOOR.PRECISE, signed steps, complete-group checks and offset bands.

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

Quick answer: FLOOR.PRECISE selects the lower multiple for positive and negative numbers, ignoring the sign of the step. =FLOOR.PRECISE(-17.2,5) is expected to return −20.

Availability and version history

  • Introduced: first launch date or introduction version not established by the official history reviewed.
  • Availability: documented in the rolling Google Sheets service; this guide uses ordinary cells in the web editor, without a claim about identical mobile authoring workflows.
  • Checked: October 14, 2026 against the official FLOOR.PRECISE reference. Examples were independently calculated; no native Sheets execution was performed.

Syntax and the default rule

=FLOOR.PRECISE(number, [significance])

The number is required. Significance defaults to 1, and Google states that its sign is ignored. With a nonzero step, the result is the greatest permitted multiple no larger than the input. There is no separate mode argument. Official FLOOR.PRECISE reference

For a negative number, a mathematical lower value is farther from zero. The function therefore differs from merely dropping a fractional part. PRECISE describes this consistent convention, not arbitrary-length numerical accuracy.

Example 1: default integer results

=FLOOR.PRECISE(17.2)
=FLOOR.PRECISE(-17.2)
=FLOOR.PRECISE(-17)

Expected results: 17, −18, −17. The value −18 is below −17.2 on the number line. The input −17 is already an integer on the default unit grid.

If a report displays −17 for an underlying −17.2 through formatting, that does not change what the rounding function receives. Inspect stored values rather than inferring them from a rounded display.

Example 2: four signed combinations

=FLOOR.PRECISE(17.2,5)
=FLOOR.PRECISE(17.2,-5)
=FLOOR.PRECISE(-17.2,5)
=FLOOR.PRECISE(-17.2,-5)

Expected results: 15, 15, −20, −20. A negative significance supplies the same five-unit spacing as a positive one.

Do not use the sign of significance as a switch for toward-zero rounding. It is explicitly ignored here. If the desired policy is different for negative records, select and document a rule that actually supports that distinction.

Example 3: distinguish integers from exact multiples

=FLOOR.PRECISE(17,5)
=FLOOR.PRECISE(20,5)

Expected results: 15 and 20. An input can have no fractional part and still need to move to a permitted grid point. This is why pack-size tests should include an integer that is not divisible by the pack size.

Advanced example: allocate complete seven-piece sets

Enter available pieces 15, 21, 22 in A2:A4. In B2 enter and fill down:

=FLOOR.PRECISE(A2,7)

Expected pieces assigned to complete sets: 14, 21, 21. In C2 use =B2/7; expected set counts are 2, 3, 3. In D2 use =A2-B2; leftovers are 1, 0, 1.

This sample allocates 56 of 58 pieces. Pooling those particular quantities also makes eight complete sets, but that agreement is not guaranteed for other inputs.

For separate quantities 6 and 6, each alone makes zero seven-piece sets. Pooling twelve pieces makes one set and leaves five. Decide whether pooling is allowed before calculating a total-based allocation.

Advanced example: lower boundaries on an offset grid

Suppose lower band boundaries begin at 1 and recur every 4 units. Use:

=1+FLOOR.PRECISE(F2-1,4)

For F2 values 2, 5, 6, −2, expected results are 1, 5, 5, −3. These belong to the sequence …, −3, 1, 5, 9, … .

With lower-inclusive and upper-exclusive bands, −2 belongs to [−3,1), 2 belongs to [1,5), and 5 belongs to [5,9). The endpoint convention is an explicit part of this example, not something a displayed boundary automatically communicates.

The subtraction and addition of the origin are both necessary. Rounding the original input and adding one afterward would implement a different grid.

Fractional steps and numerical checks

=FLOOR.PRECISE(9.1,0.25)
=FLOOR.PRECISE(-9.1,0.25)

Expected results: 9 and −9.25. Both are quarter-unit multiples and no larger than their corresponding inputs.

For ordinary exact arithmetic and a positive step, the chosen lower point should be at most the input, while the next higher grid point should exceed an off-grid input. These checks help diagnose a mistaken step or offset. Measurements very close to boundaries need actual workbook testing with an appropriate numerical tolerance.

INT supplies a simple integer floor. ROUNDDOWN uses decimal places and rounds toward zero; it is not an equivalent negative-value rule.

Keep the unrounded source when later calculations need its original precision. The checked Google reference does not give a complete zero-step, text-coercion or extreme-value error table, so this guide uses nonzero numeric steps rather than copying another spreadsheet’s boundary behavior.

Verification: Expected results use fictional data and independent calculations rather than native Google Sheets observations. See the editorial testing standards.