PERCENTILE Function (LibreOffice Calc)

PERCENTILE finds the value at a chosen fractional position in a data distribution. Use it for statements such as “the 90th-percentile processing time” rather than claiming it returns a count of items. Its second argument, Alpha, is a number between 0 and 1 (for example, 0.9 for the 90th percentile).

=PERCENTILE(A2:A6;0.9)

With observations 4, 8, 12, 16, 20, LibreOffice Calc 25.2.3.2 returns 18.4. This is an interpolated value and need not appear among your original observations.

PERCENTILE(Data; Alpha)

Data — Numeric data range or array; text and empty cells in referenced ranges are ignored. A2:A6

Alpha — Fractional percentile from 0 (minimum) through 1 (maximum). 0.9

Calc uses semicolons by default in many installations. Using 90 instead of 0.9 is a common mistake: it requests a position beyond 100% rather than the 90th percentile.

Open the practice CSV and keep column A as numeric values:

On a narrow screen, scroll this table sideways to see every column.

Enter the two formulas in separate empty cells:

=PERCENTILE(A2:A6;0.25)
=PERCENTILE(A2:A6;0.9)

Native-tested results: 8 (25th percentile) and 18.4 (90th percentile).

Why is the 90th percentile 18.4?

For the inclusive, linearly interpolated scale used by this function, sort the five values and calculate the rank as 1 + (5 - 1) × 0.9 = 4.6. The answer lies 60% of the way from the fourth value (16) to the fifth (20): 16 + 0.6 × (20 - 16) = 18.4.

This shows why a percentile isn’t necessarily a measurement actually recorded. In business reports, document the percentile method when sharing results between tools.

=PERCENTILE(A2:A6;0)
=PERCENTILE(A2:A6;1)

Native-tested results: 4 and 20. The median would be the 50th percentile, here 12. Percentiles do not by themselves tell you how many rows exceeded an operational target; use COUNTIF for that question.

In this dataset:

On a narrow screen, scroll this table sideways to see every column.

The inclusive forms accept the endpoints 0 and 1; the exclusive form does not. LibreOffice Help documents the .INC and .EXC functions as available since LibreOffice 4.3. Don’t switch methods in an existing analytics dashboard without checking how the benchmark changes.

Suppose column A contains actual service times in minutes. A 90th-percentile value can help evaluate whether long-tail delays are improving. Save the calculation definition, sample window and number of valid observations. A single new high observation can change the percentile; small sample sizes warrant careful interpretation.

To label results, enter a neighbouring text label such as 90th percentile (min) instead of embedding measurement units into your formula, which could convert the result into text and interfere with later calculations.

• Wrong Alpha: use 0.9, not 90 for 90%.
• Numbers treated as text: PERCENTILE ignores text in a referenced range, potentially reducing the true sample size.
• Unexpected decimal answer: interpolation can produce values not present in the original data.
• Method mismatch: percentile numbers from .EXC and .INC need not agree, even on identical data.
• Changing range: if new rows are appended outside A2:A6, extend the reference or use a suitable maintained range.

This page describes Calc’s implementation. Excel also uses PERCENTILE, PERCENTILE.INC, and PERCENTILE.EXC, but separators and compatibility settings can vary. Validate the chosen method when moving reports between applications.

The five PERCENTILE calculations above were run in LibreOffice Calc 25.2.3.2 on October 10, 2026 with no formula errors. No native testing of other spreadsheet programs is claimed.

• Official Calc PERCENTILE, .INC and .EXC documentation.
• MEDIAN, AVERAGE, MIN and MAX.

Selected example formulas were run in LibreOffice Calc 25.2.3.2 on Linux on October 12, 2026. The official Help may cover a newer version. Tests of selected formulas do not certify every locale, interactive array operation or file round trip. View the example-check record.

Official PERCENTILE reference · Editorial standards