AVERAGEA in Collabora Office: Average Numbers While Counting Text as Zero

• 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: online and mobile deployments are not inferred from desktop documentation.
• Checked: October 14, 2026. Function-specific documentation and independent calculations; no native Collabora test was performed.

=AVERAGEA(argument1;[argument2];...)

Collabora documents up to 255 arguments, which may be numbers, text, cell references or ranges. That argument limit is not a claim that a range may contain only 255 cells. The examples use English function names and semicolon separators.

The critical decision is whether text really should count as zero for the question being answered. A word such as Pending is not a measured zero, even though this function assigns it that contribution. Preserve the original label and explain the reporting policy instead of silently converting missing observations into apparent measurements.

Enter this small register in A1:B5:

=AVERAGEA(B2:B5)
=AVERAGE(B2:B5)

Expected results are 15 and 20. AVERAGEA uses (12+18+30+0)/4; AVERAGE excludes the text from this range and uses 60/3. Both results are consistent with their definitions, but they answer different inclusion policies.

For an explicit scalar control, =AVERAGEA(12;18;30;"Pending") also has expected result 15 under the documented text-as-zero rule. The text is deliberately ordinary nonnumeric text, avoiding an unverified numeric-text conversion case.

Add another text Pending in B6 and extend the range to B2:B6. The expected AVERAGEA becomes 12, because the numeric total remains 60 while the participating count becomes five.

Changing the first Pending to numeric 15, while leaving the second Pending, gives 15, from (12+18+30+15+0)/5. Removing B6 from the calculation instead gives 18.75, from the four numeric measurements. State which operation was performed; extending a range, entering a value and excluding a record are different edits.

Create these two separate groups:

• A2:A4: numeric 8, numeric 12, text Pending
• C2:C3: numeric 30, text Pending

=AVERAGEA(A2:A4)
=AVERAGEA(C2:C3)
=AVERAGEA(A2:A4;C2:C3)

Expected results are approximately 6.666666667, 15 and 10. The combined calculation has five participating values and total 50.

A simple average of the two group results is approximately 10.833333333, which gives equal weight to groups of different sizes. The correct weighted reconstruction for these explicit controls is:

=(AVERAGEA(A2:A4)*3+AVERAGEA(C2:C3)*2)/5

Expected result: 10. The counts three and two include the text entries because that is the chosen AVERAGEA policy. Using only the numeric counts would not reconstruct this mean.

For two cells containing ordinary text Pending and Review, the documented rule gives an independently expected mean of 0. That does not prove either item has a measured score of zero. Retain a separate completion status rather than treating the aggregate number as evidence that all observations arrived.

A compact report can show the total expected items, actual numeric observations and the chosen average side by side. COUNT helps inspect numeric observations in these controlled ranges, but does not by itself explain every missing-data policy.

The function-specific source is concise. These examples establish ordinary numeric and nonnumeric text controls; they do not certify a complete Boolean, formula-empty-string, numeric-text, error-containing or entirely empty-range matrix. In particular, a cell with no content is not silently treated as equivalent to the supplied text Pending here.

If the answer is lower than expected, inspect whether a text label has increased the denominator. If it changes after an import, check the stored types before changing the formula. Keep the full-precision result for further arithmetic and round only the displayed report when appropriate.

• Collabora Office 26.04: AVERAGEA.
• Editorial testing standards.

The function-specific Collabora documentation was reviewed. Worked results are independent arithmetic or documented type-rule controls, not captured output from the application. Client-specific entry procedures and the boundaries identified above remain subject to native verification.