Zoho Sheet DAVERAGE Function: Selected Means and Weights

Zoho’s official DAVERAGE documentation was retrieved on October 15, 2026. It documents the function in the rolling web service. The first release and last functioning version are unknown; this is not a historical build test. Mobile/offline parity is unverified. Results below are independently worked expectations, not captured Zoho Sheet execution. The datasets are fictional and fully specified.

DAVERAGE(database; field; criteria)

All three arguments are documented. database includes a header row and its records. field identifies the column to process, using either its quoted header text or a reference to that header cell. criteria points to a separate range containing the relevant header and condition. The condition column and the result field can be different.

Use semicolons as shown in Zoho’s English reference; adjust separators if the workbook locale requires it. A header reference such as E1 is unquoted, whereas a literal name such as “Amount” needs quotes. These examples use explicit field names, not an assumed numeric field index or omitted argument.

Create a sheet named Jobs. Enter the following data in A1:F7. Headers are literal text; Units, Rate and Duration are numeric. In E2 enter =C2*D2 and fill through E7 to generate Amount. The displayed amounts below are checks, not extra inputs.

Set H1 to Team and H2 to Red. Set J1 to Team and J2 to Blue. These are two separate criteria ranges, H1:H2 and J1:J2. Red selects rows 2,4,5,7; Blue selects rows 3,6. Use the formulas on Jobs so the unqualified references point to this dataset.

The amount mean is (20+60+20+30)/4. Do not divide by all six database records after filtering to Red. With L1=Team and L2 physically blank, =DAVERAGE(A1:F7;"Amount";L1:L2) gives approximately 34.333333333, the mean of all six amounts under the documented blank-below-header rule.

In a copy, change E4 from its formula to numeric 100 for an explicit what-if scenario. Red’s amount mean becomes 42.5; Blue remains 38. This replacement deliberately breaks that row’s Units×Rate link. Label the copy as a scenario so it is not confused with the original derived-amount data.

Red has four records and Blue has two. Averaging their means gives (32.5+38)/2 = 35.25, which gives the two teams equal weight. The six-record mean is 34.333333333. To reconstruct it from group summaries, weight the means by their record counts:

=(32.5*4+38*2)/(4+2)

This weighting is valid for this clean numeric field because the counts represent the actual observations in each mean. If another field has missing or nonnumeric values, count the participating values for that field rather than automatically reusing the number of records.

The mean of Red’s four rates is 13.75. Its amount divided by units is 130/10 = 13. Use =DSUM(A1:F7;"Amount";H1:H2)/DSUM(A1:F7;"Units";H1:H2) for that latter quantity. DAVERAGE does not automatically apply weights from the neighboring Units column.

If every quantity were one, the two rate summaries would agree. In the supplied table they do not, because the 15-rate row represents four units while the 20-rate row represents one. Choose the statistic that matches the question, and make its denominator explicit.

A numeric zero is a valid observation, not missing data. Replacing E4 with zero in a separate copy gives Red (20+0+20+30)/4 = 17.5. Deleting the value and entering text are different inputs whose detailed coercion behavior is not established here.

The reference associates #DIV/0! with an empty criteria range, yet separately permits a blank condition below a valid header for whole-column processing. Preserve that distinction. A mean with no participating numeric observations has no mathematical value; this guide does not claim an independently observed error token for every no-match or malformed-criteria case.

The official DAVERAGE page lists shared name, type and reference errors. It does not exhaustively specify no-match behavior, mixed-type coercion, wildcard rules, duplicate header handling or every exact error token. Do not import those details from another spreadsheet product simply because the function name matches.

Start with the small dataset above. Confirm that the database includes row 1, the criteria header matches the intended field and all references point to the intended sheet. Inspect underlying cell contents rather than display formatting. Keep criteria outside the source records so editing a condition does not alter the database. A correct formula can still answer the wrong question if its range omits newly appended rows.

Before transferring a workbook, compare these controls in the destination app. This guide does not establish Excel, Numbers or other products’ syntax, error or version equivalence.

DSUM and DCOUNT help expose the numerator and numeric-observation count. A shared function name across products does not prove identical treatment of blanks or text.